11.5V AC Dimmer

Go To Last Post
28 posts / 0 new
Author
Message
#1
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Hi,
I have searched around and googled for this but couldnt quite find what I need.
I want to build a 8 channel dimmer for 11.5V (12V) AC halogen lamps. The chanels have to be controled with an AVR so that I can programm the outputs.
The way to do this is simple with a 8 chanel PWM. But I do not know how to control the AC Voltage for the Lamp. I thought of a small circuit for each lamp using a Triac. But how can I combine the 11.5V AC for the Lamps with the 5V DC for the MC?
Thanks for any help.
Regards

Attachment(s): 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I don't know for sure but do you need AC?
...why not use DC, it would be easier to control.

Regards
heguli

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Would'nt a FET be the way to do this ???

I mean it's just 12v

/Bingo

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Bingo600 wrote:
Would'nt a FET be the way to do this ???

I mean it's just 12v

/Bingo

Not for AC, as the reverse biased diode integral to the part would conduct through 1/2 of the cycle. You can use a FET or transistor to drive the Triac.

Writing code is like having sex.... make one little mistake, and you're supporting it for life.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I'm a little concerned about whether a triac will function correctly at such low voltages. I have no experience one way or the other, just a concern.

The lamps are not discharge devices. I would bet that they will work just fine from DC. Halogen headlight bulbs for automobiles certainly do. It would make things a lot easier. On the other hand, there just MIGHT be some design characteristic that targets them specially for AC. Again, no experience, one way or the other.

With a triac, yes, you might want to use PWM, but it would need to be synchronized with the line phase. You can't (with good results, at least) just PWM independent of the line signal. This would take it out of the usual counter-based PWM and would probably be slow enough to do it all in software.

Jim

 

Until Black Lives Matter, we do not have "All Lives Matter"!

 

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

If you really need AC, why not just use an H-bridge. Using an H-bridge, wouldn't the Halogen bulb be seeing the equivelent of an AC voltage? The size of the H-bridge would be determined by the current draw of the Halogen bulb. For something less then 5 amperes each, an LMD18201T would be nearly a perfect fit for your needs.

But I agree with others that have posted above, I really don't think you need an AC source to drive a Halogen bulb.

Just keep the oils from your fingers off of the glass, else they go "Poof"!

Edit:
Isn't this H-bridge idea of mine just a huge "Brain Fart?" A plain old power MOS-FET would work!

You can avoid reality, for a while.  But you can't avoid the consequences of reality! - C.W. Livingston

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

With an H-bridge, you could certainly generate the same effect as AC. Not such a bad idea. The direction control pin would control the "polarity".

Jim

 

Until Black Lives Matter, we do not have "All Lives Matter"!

 

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I would expect that the reason my Halogen Spots runs 12v AC is that they save the Diodebridge in the Trafo Box.

There are 2 ways to control this , but a lot of times i see "Inductive dimmers for Halogen" mentioned ... meaning that they control the 220v mains going into the trafo , not the 12v comming out.

/Bingo

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

The diode bridge would also generate quite a bit of heat.

Writing code is like having sex.... make one little mistake, and you're supporting it for life.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

diaforic wrote:
Hi,
I have searched around and googled for this but couldnt quite find what I need.
I want to build a 8 channel dimmer for 11.5V (12V) AC halogen lamps. The chanels have to be controled with an AVR so that I can programm the outputs.
Regards

Greetings,

Let's take a step back. What power source do you have now? Are the 11.5V (I'll call them "12V" for simplicity) lamps running from the AC mains through a transformer? Is there one transformer per lamp or are they on a bus connected to a single transformer? There are a lot of modern architectural lamp fixtures that have multiple low voltage bulbs and exposed bus bars connected to a single transformer.

If you are starting from scratch the easiest method is to use one PMOS (PowerMOS) transistor per lamp, and operate all the lamps from a common 12V dc power supply, that in turn is connected to the AC mains. Each PMOS device would use duty-cycle (PWM) to vary the average current in the lamps and hence its brightness. Driving the PMOS devices is trivial and can be done directly from the uC port.

On the other hand, if each lamp has its own transformer it might make more sense to control the AC to the transformer with a Triac. This is how the standard wall switch dimmers works. As long as the waveform to the transformer is symmetrical the lamp will follow the average of the input voltage and no harm will be done to the transformer. Instead of PWM the Triacs are 'fired' at varying times after the AC waveform crosses zero, and the Triac remains conducting until the next AC zero-crossing. An early trigger gives almost full power to the load (complete AC sinewave) and late trigger gives progressively less power to the load. This is used in home and theatre lighting dimmers. A sync circuit to detect the AC zero crossing is required, and could easily feed the uC interrupt.

The down side to doing this that driving each Triac requires an interface circuit. A simple transistor stage and local 12V (or more) DC supply could be used. The uC ground will be common the AC mains neutral and care must be taken to make the finished project safe. Obviously, the uC could not be connected to earth ground and could not allow data com (common Serial or Parallel connections) or other interconnections.

Finally, the problem could be solved with an opto isolator that connects directly to both the Triac and the uC port, but provides safe isolation of the AC and low voltage DC (ground referred) sides.

If there isn't a transformer-per-lamp today it would be tempting to use a common transformer and place the Triacs on the secondary side, one per lamp. This would still require a driver stage between the Triacs and the uC.

In all cases their will be some power lost in the electronic switches. For a Triac this is approximately 1.2V and would have much less effect in the AC mains side than the low voltage side of the transformer. PMOS devices can be found with very low Ron (ON resistance) and would win the greater efficiency award.

Comments Welcome!

--
Peter J. Stonard
www.stonard.com

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Most circuits that you search on for this application rectify the 12 VAC and then chop with a PWM. At these low voltages, a triac is problematic, and you have to determine zero crossing.

Arcs_n_Sparks

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

ka7ehk wrote:
With an H-bridge, you could certainly generate the same effect as AC. Not such a bad idea. The direction control pin would control the "polarity".

Jim

Yeah but, after I thought about it, using the H-bridge in the conventional manner the minimum power (and hence the minimum light intensity) would be at the fifty percent duty cycle point. Anything above or below 50% duty cycle would tend to increase the light intensity as the duty cycle approached 100% or zero percent duty cycle.

And that was the "Brain Fart!"

You can avoid reality, for a while.  But you can't avoid the consequences of reality! - C.W. Livingston

Last Edited: Thu. Apr 13, 2006 - 06:40 AM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I've seen a couple of "Light Electronic Transformers" that are provided specifically for 12V lighting purposes. They are very light, contain very little iron, operate similar to a switching power supply except the output is not rectified or filtered. I expect the 12V AC output frequency to be somewhere in the inaudible KHz region. I believe that these "Transformers" can be fed from conventional lighting dimmers.

Just throwing another wrench into the works...

Stan

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Hi,

it´s sayd a lot using DC. It has advantages for sure, but also disadvantages.

The OP wants to use AC. So i try to look for an AC solution.

First i rtecommend using an xtal for your ucontroller to have a reliable timing.
Then - because you might not have 8 hardware PWM - you should craete 8 software PWM.

this must be synchronized with mains frequency. Zero crossing detection. every time (falling and rising) you detect a zero crossing start a counter for the PWM.
The counter frequency should be mains frequency x 2 x 32. (32 is for the number of steps - i think 32 steps are enough)

for bright light start your "ignition" with a short delay (with upcounting counter use low compare values)
for dark light use a long dely.
switch OFF PWM BEFORE you expect the zero cross. -- At high counter values (> 27 for example) to prevent a RESTART at the wrong halfwave.

There are spacial triacs with (maybe isolated) DC input. I recommend this ones. The input behaves like a LED - the output like a triac.

Good luck

Klaus
********************************
Look at: www.megausb.de (German)
********************************

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

You will try these schematic

Attachment(s): 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

With such a low lamp supply voltage, you may just want to eliminate R3, reduce
R5 to just a few ohms- maybe 2.2 to 4.7, and snubber R6 & C2 won't be needed at all.

edit: and you'll want a "random turn on" opto instead of a "zero cross"

Tom Pappano
Tulsa, Oklahoma

Tom Pappano
Tulsa, Oklahoma

Last Edited: Thu. Apr 13, 2006 - 10:44 AM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Ok. Thank you all very much for your help. :D That should give me enough Info.
I will try driving the lamps with DC and then see if it works. If not I have to fiddle somethig and try the circuit postet last.
If anyone is interested I will post my results sometime in the futher future on my website.
Best Regards

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

And to remind, without capacitor after rectifier max. voltage drops 1,4 V. Some capacitors might be in order to get 12 VDC.

Regards
heguli

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

sorry to not be agree with all people but DC you will lost a lost of power in the bridge.
AC will be a nive solution with a triac. You can do must more easyere than the solution with opto coupler.
But it is right you must have a good Zero crossing detection to have a stabilt ligth.
I have done a lot of this system for disco with 0-10V drive but you can do much easyere with direct drive from CPU.
If you want some help for AC & Triac I can help you. See me a mail.
Thierry

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Hi

@tappano:

Quote:
snubber R6 & C2 won't be needed at all.

I´m not sure, but here my thoughts:
Here not the load is inductive, but the transformer will have some inductance. (also wiring to load an transformer)
Doesn´t the snubber help for the transformer´s inductance also?
Can´t the snubber help to prevent form high dU/dt in common?

Klaus
********************************
Look at: www.megausb.de (German)
********************************

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Have an uneasy feeling about the circuit above that it can only be use with resistive load.
I wouldn't use a transformer for this circuit.
I have done it once & never again because of the sharp rise time energy was enough to break down the insulation of the primary winding. It short circuited itself.

Pretty sure there is a different layout like using for a ceiling fan speed controller.

Ken

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Quote:
I´m not sure, but here my thoughts:
Here not the load is inductive, but the transformer will have some inductance. (also wiring to load an transformer)
Doesn´t the snubber help for the transformer´s inductance also?
Can´t the snubber help to prevent form high dU/dt in common?

Good point Klaus,

The transformer could look like a small inductor due to leakage inductance.
The wiring, if imbalanced, would contribute a small amount too.

Quote:
Have an uneasy feeling about the circuit above that it can only be use with resistive load.
I wouldn't use a transformer for this circuit.
I have done it once & never again because of the sharp rise time energy was enough to break down the insulation of the primary winding. It short circuited itself.

I'm curious what your actual setup was that failed. I'm using thyristors on
motors at 480 volts.

Tom Pappano
Tulsa, Oklahoma

Tom Pappano
Tulsa, Oklahoma

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

glitch wrote:
Bingo600 wrote:
Would'nt a FET be the way to do this ???

I mean it's just 12v

/Bingo

Not for AC, as the reverse biased diode integral to the part would conduct through 1/2 of the cycle. You can use a FET or transistor to drive the Triac.

You can get around the problem wih the integral bulk diode by placing FETS face-to-face.

DFR

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

ITP wrote:
sorry to not be agree with all people but DC you will lost a lost of power in the bridge.
.....
Thierry

Greetings,

Please explain what you mean by "lost of power". Have you calculated the efficiency of a similar circuit? Do you have any theory to support your claim?

Thanks In Advance!

Peter

--
Peter J. Stonard
www.stonard.com

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

danafraymond wrote:
glitch wrote:
Bingo600 wrote:
Would'nt a FET
You can get around the problem wih the integral bulk diode by placing FETS face-to-face.

DFR

Greetings,

True, but doesn't that double the number of FETs required? This is an eight channel project, so wouldn't a single full-wave bridge rectifier be less costly than eight more PMOS FETs? Wouldn't extra FETs also complicate the driver stage(s)?

By face-to-face do you mean in series? Two PMOS FETs in parallel would not prevent the conduction of (either) body diode. Placing a suitable power diode in series with a single PMOS FET will block reverse conduction through the PMOS FET.

A PMOS FET with diode would only conduct on one half cycle of the AC, thus limiting the load to 50% power at maximum. The AC input voltage could be raised to 23V RMS (11.5 *2) so that at the 100% level the lamps see 23V for 50% duty cycle and produce an average output the same as if run from DC at 11.5V.

On the other hand, a full wave bridge (without a capacitor) would work well with a single PMOS FET (per lamp load), even though the rectified AC is not filtered. Large and costly electrolytic capacitors would be avoided. In this case the uC PWM would require synchronization to the AC zero-crossing to prevent alias (beating) of the PWM and AC signals (which might create objectionable flicker in the lamps).

Comments Welcome!

Peter

--
Peter J. Stonard
www.stonard.com

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

The power loss is in the diode bridge. The voltage drop that a diode creates on the input voltage is converted to heat. Assuming a 1V drop per diode, and a 50W load (~4A@12v, ~400ma@120V) means that the bridge will be generating 8W of heat (that's quite a bit), if done at 12V. On the other hand, if we place the bridge on the 120V side, the power loss is 1/10th since the current is 1/10th. Unfortunately transformers don't work well on DC.

Writing code is like having sex.... make one little mistake, and you're supporting it for life.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

glitch wrote:
The power loss is in the diode bridge. The voltage drop that a diode creates on the input voltage is converted to heat. Assuming a 1V drop per diode, and a 50W load (~4A@12v, ~400ma@120V) means that the bridge will be generating 8W of heat (that's quite a bit), if done at 12V. On the other hand, if we place the bridge on the 120V side, the power loss is 1/10th since the current is 1/10th. Unfortunately transformers don't work well on DC.

Glitch,

Thanks for the detail. Are you also user ITP (Theirry), by chance? I was interested in his specific post on this project.

The expected power loss in the bridge could be reduced by the use of lower voltage drop rectifiers. For example:
http://tinyurl.com/nxk89

With a max Vf of 0.35V the same fifty watt load at 12V would produce only a 3.9W loss, or stated another way, a 65% improvement over your example. This brings the DC power supply up to approximately 92% efficiency.

While on the topic of efficiency, with eight channels (50/8 = 6.25W) what power is lost in the PMOS FET or Triac?

Using a low RDSon PMOS FET. For example:
http://tinyurl.com/q53fb

The power lost (per channel) is likely to be around 50mW, or 400mW total.
Heatsinks would not be required.

Using a Triac. For example:
http://tinyurl.com/j5dnx

The power loss (per channel) is likely to be around 800mW, or 6.6W total.
Heatsinks would not be required.

So as I see it the extra losses of using a bridge rectifier (and low Vf diodes) in combination with PMOS FETs all on a single transformer is the most efficient (and therefore reliable) solution. Plus, as noted earlier, the interface to a uC is simple with PMOS FETs.

We are waiting to hear back from the OP with details on whether there is to be one large transformer or eight smaller ones, one per lamp. In fact, as yet we don't know the power rating of each lamp.

Comments Welcome!

Peter

--
Peter J. Stonard
www.stonard.com

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I never said FET's weren't a good option, in fact they can be. My only comment was the loss on your typical power diode, with good design yes you can make an efficient power supply.

And no I'm not ITP, this is my only identity here.

Writing code is like having sex.... make one little mistake, and you're supporting it for life.