Is it possible to build a Variac without using Transformers?

[MMDuino]

In my works with High voltage (220VAc-50Hz) always i need to (variable AC voltages), so my inquiry is; are there anyone made variac without variable transformers?, for example by using Thyristors, Triacs or even Transistors and get on adjustable 220 vAc with normal sine wave in the result?

Thanks in advance

[Simon]

simply no, if you want to do it with electronics parts your looking at making your own variable output inverter or buying one, very complicated and expensive.

The real variac is the simplest and cheapest solution.

[T3sl4co1l]

You can build a buck converter using synchronous switching and bidirectional switches, which implements the same idea as its DC counterpart, but works for AC as well.  At a fixed PWM%, it will behave much the same as a variac (equivalent winding resistance and everything).

It most certainly will not have the surge and fault capabilities of a real variac, though.

Tim

[Simon]

You would need a dual polarity buck converter with a varying output.

[robrenz]

You can do it with a bipolar operational power supply as shown in this thread.  You can drive it with a signal generator to get the frequency and amplitude you want. You would need a 500V unit to do the 220VRMS.  Very handy piece of equipment in general.

[SeanB]

If you do not mind having discrete steps you can just use a transformer with multiple taps, and select them as needed, using either a switch or relays. You can have one wound with a series of taps giving you from 60VAC ( really a good place to start) to 250VAC, using taps every 10V on the transformer. 21 wires later and you can go from 60 to 250VAC using some switching.

[Richard Crowley]

Yes, it can be done, but it is a very dangerous and complex circuit that many people with decades of experience (including me) would never attempt to make.
If you want a real sine wave output, you are looking at an enormous linear power amplifier, not something that can be built with thyristors.

Of course it would never give you the ISOLATION of a true transformer. But then many variacs don't provide isolation either because they are simply auto-transformers.

[Simon]

If you do not mind having discrete steps you can just use a transformer with multiple taps, and select them as needed, using either a switch or relays. You can have one wound with a series of taps giving you from 60VAC ( really a good place to start) to 250VAC, using taps every 10V on the transformer. 21 wires later and you can go from 60 to 250VAC using some switching.

The whole point apparently is to avoid the transformer. 

[Jeroen3]

I've seen a variable supply, for automated testing purposes, using a normal transformer and a thyristor dimmer on the primary.
Depending on the required quality and power, you can achieve the same using stepped approximation, such as an UPS. Or maybe even dirty pwm like a motor drive. You might need some filtering with the last one.

[German_EE]

It should be possible but I would not like to attempt the task.

1) Rectify and smooth the incoming mains supply to DC.

2) Pulse width modulate the DC to build a 50/60 Hz waveform at the wanted voltage.

3) Put your output through a low-pass filter which will hopefully remove as much as possible of the RF hash and other rubbish that you will see on the output and leave you with a reasonable sine wave.

[SeanB]

The whole point apparently is to avoid the transformer. 

No, to avoid the variable transformer. I had a transformer I got as surplus once which had a mains primary and secondaries of 20VAC, it just had 20 of them. Thus you could make almost any output from 20VAC to 400VAC just by wiring up the secondaries.

[albert22]

Here is a solution for a small current "electronic variac"
http://web.tiscali.it/i2viu/electronic/variac.htm
IMO this is the simplest solution for low power applications.

[T3sl4co1l]

Here is a solution for a small current "electronic variac"
http://web.tiscali.it/i2viu/electronic/variac.htm
IMO this is the simplest solution for low power applications.

The first one is as good as any old rheostat (protip: you can use one of those here, too, in place of the transistor stuffs hanging on the bridge), with the difference that you need big heatsinks on relatively cheap transistors, instead of some bigass ceramic and metal monstrosity.

The second one is not, however, because it applies a constant current -- the output voltage might still be in regulation, but it won't be sinusoidal.  An OTA could be used to multiply the feedback signal by the bridge voltage, implementing an electrically variable resistor.  Then both output regulation (active AC regulation, cool!) and sinusoidal response (well, for the resistor part anyway) will be had.

Tim

[Dave]

Well, you could rectify mains directly and use PWM and some nice fat IGBTs to generate the output waveform (well filtered, of course). Kind of like a variable frequency drive.

But that would be a very tricky and potentially dangerous design job. Not something a beginner should attempt.

[Simon]

Dear All

Thank you very much, really i learned good directions from the points which were raised here, i will collect all of your points then i'll try to get good result.
I will try today to do (555 PWM or 2 Transistors as multi vibrator to generate 50/60 Hz & "IGBT MOSFET HGTP10N50F1D" as a High voltage driver) and i'll connect Bulb 200 watt 220vAC-50Hz on the output to catch the voltage on the output side, but really i dont know if the result will be normal sine wave cause unfortunately i don't have Oscilloscope....

That won't work, you will get a square wave and you need to use a H-Bridge output to get negative and positive voltages that make up AC, as we keep trying to tell you this is not simple. How will you vary your voltage ? you can't.

The way it works if you do it electronically is you use a high frequency PWM much higher than 50Hz and you use different PWM duties to simulate the varying voltage of the AC since wave so that you can approximate a sine wave, you can then also filter it to make it better but this is all very complicated.

With your level of knowledge you should not be playing with the mains unless your qualified in mains work like being an electrician.

[MMDuino]

Oh, really i forgot that, the result will be square wave with PWM,,
ok, i'll change my direction by using Sine wave generator and transistor 2SC3322 (Cause i have some pieces of it)with 10k variable resistor on the base to control the output of the transistor..
do you agree with me?

please see the next sin wave generators...

Dear All

Thank you very much, really i learned good directions from the points which were raised here, i will collect all of your points then i'll try to get good result.
I will try today to do (555 PWM or 2 Transistors as multi vibrator to generate 50/60 Hz & "IGBT MOSFET HGTP10N50F1D" as a High voltage driver) and i'll connect Bulb 200 watt 220vAC-50Hz on the output to catch the voltage on the output side, but really i dont know if the result will be normal sine wave cause unfortunately i don't have Oscilloscope....

That won't work, you will get a square wave and you need to use a H-Bridge output to get negative and positive voltages that make up AC, as we keep trying to tell you this is not simple. How will you vary your voltage ? you can't.

The way it works if you do it electronically is you use a high frequency PWM much higher than 50Hz and you use different PWM duties to simulate the varying voltage of the AC since wave so that you can approximate a sine wave, you can then also filter it to make it better but this is all very complicated.

With your level of knowledge you should not be playing with the mains unless your qualified in mains work like being an electrician.

[Simon]

It's not that simple, your transistors control current not voltage so you need to vary the average amplitude of the drive voltage from your oscillator and have feedback from the output and you need to work in both negative and positive voltages.

Again this really is not a simple thing to do and not advisable for you to try for quite some time.

[macboy]

Oh, really i forgot that, the result will be square wave with PWM,,
ok, i'll change my direction by using Sine wave generator and transistor 2SC3322 (Cause i have some pieces of it)with 10k variable resistor on the base to control the output of the transistor..
do you agree with me?

No.

[German_EE]

Don't forget that if you want to feed a sine wave into a power transistor to get a bigger sine wave out then the transistor will need to be biased into the Class A linear region. This means that the efficiency will be very bad, somewhere between 33% and 25%, so for 100W of power in you will generate over 60W of heat.

[Dave]

But that would be a very tricky and potentially dangerous design job. Not something a beginner should attempt.

Have I not stressed this enough?

I believe the problem here is that you are not even aware of how little you know about the subject. While this is often very good, as it encourages learning, this is not the case here. Mains voltage itself is very dangerous and you would be rectifying it and charging a capacitor with it, making it even more lethal. One small mistake and you are dead.

Buy a proper variac and forget about this thing for now. Start off with something easier, something that can't kill you.

[Dago]

Oh, really i forgot that, the result will be square wave with PWM,,
ok, i'll change my direction by using Sine wave generator and transistor 2SC3322 (Cause i have some pieces of it)with 10k variable resistor on the base to control the output of the transistor..
do you agree with me?

please see the next sin wave generators...

Dear All

Thank you very much, really i learned good directions from the points which were raised here, i will collect all of your points then i'll try to get good result.
I will try today to do (555 PWM or 2 Transistors as multi vibrator to generate 50/60 Hz & "IGBT MOSFET HGTP10N50F1D" as a High voltage driver) and i'll connect Bulb 200 watt 220vAC-50Hz on the output to catch the voltage on the output side, but really i dont know if the result will be normal sine wave cause unfortunately i don't have Oscilloscope....

That won't work, you will get a square wave and you need to use a H-Bridge output to get negative and positive voltages that make up AC, as we keep trying to tell you this is not simple. How will you vary your voltage ? you can't.

The way it works if you do it electronically is you use a high frequency PWM much higher than 50Hz and you use different PWM duties to simulate the varying voltage of the AC since wave so that you can approximate a sine wave, you can then also filter it to make it better but this is all very complicated.

With your level of knowledge you should not be playing with the mains unless your qualified in mains work like being an electrician.

No it will not work. You need a sine-wave modulated square wave (with dead-time) and you need to filter the transformer output to get a sine-wave.

Like others have said, this is not an easy thing to design even for an experienced designer, not a suitable project for a beginner.

[iamdarkyoshi]

Just do what my dad did and put a sine wave into an protected audio amp.

[Simon]

judging by the lack of responses from the op he has already killed himself.

[MMDuino]

Thanks for your care, but i'm still here breathing....

that's mean you encourage me to tear down my UPS and use it's transformer to get good results..

ok. i will try that, then i will post my results...

judging by the lack of responses from the op he has already killed himself.

[Simon]

Thanks for your care, but i'm still here breathing....

that's mean you encourage me to tear down my UPS and use it's transformer to get good results..

ok. i will try that, then i will post my results...

judging by the lack of responses from the op he has already killed himself.

and what makes you think a UPS transformer will help ? it may be totally the wrong type as it could be a high frequncy step up followed by a high voltage AC generation circuit.

[rickselectricalprojects]

I brought a "scr dimmer" of ebay and it could apparently vary ac voltage from 10-240 v ac. I don't remember how i wired it up but i Ballested it with two heaters and put a resistive load (a lightbulb). I did get a output voltage of 10-240v but it DID NOT have a normal sinewave. The only thing that was good for was blowing up speakers and it still was very bad at that. So the best bet would be to go and pick up a cheap and/or used variac of ebay

[Richard Crowley]

I did get a output voltage of 10-240v ....

No, you didn't really.  You got the 240 VAC RMS sine wave chopped up to vary the average power to the resistive load with a good amount of "averaging" in the thermal mass of the incandescent lamp filament.

[Refrigerator]

Well, if you put 350 diodes across your 240V_AC you can use the diode voltage drop to chop the voltage up...

[madires]

I did get a output voltage of 10-240v ....

No, you didn't really.  You got the 240 VAC RMS sine wave chopped up to vary the average power to the resistive load with a good amount of "averaging" in the thermal mass of the incandescent lamp filament.

Some references including nice pictures and graphs:
- http://www.ilight.co.uk/downloads/iLIGHT%20Binder-HowDimmers.pdf
- http://home.howstuffworks.com/dimmer-switch.htm
- https://en.wikipedia.org/wiki/Dimmer