upgrading 110/220V Ac controller to DC

[Trickyd]

Hello Light fans,

As AC controllers get more out of use these days, I got the strange idea to upgrade the high voltage AC sides to low voltage DC.

Not knowing the exact schematics, but with a little knowledge of electronics, I came to the idea to replace the triacs by fets and adjusting  the seconday circuit and so on. while uncoupling the 220v input of the transformer from the power input and giving its own 220V entry.

As the circuit was driving the triacs every half phase, this should result in a frequency of 100 HZ ( in europe ) driving the fets. with some kind of pwm.

 

Has anyone tried this? If so, are you willing to share the new secondary schematics? Please feel free to comment. I was just thinking it is a waste to throw the AC controlers in the recycle bin.

 

Dick de Wit

The Netherlands

[k6ccc]

That seems like an awful lot of work for a fairly small gain. Without knowing the details of the AC controllers design, it is likely possible, but would take a lot of parts changes and board surgery. Not worth my time considering the price of the DC controllers.

Sent from my Droid Turbo via Tapatalk

[a31ford]

You would have to replace the optocouplers to a zero crossing type and as far as i know simply a heavy drive transistor (MOSFET) would do and even larger heat sinks, uncouple the units power transformer from the triac suppy rails and it might be a good idea to beef-up the foil on the supply rails since you will be pushing alot more current at a lower DC voltage.

ABOVE ALL, i would try a opto/drive transistor OFF board, from a single channel FIRST ! Remember that the output side of the opto is live to the return rail on the board. Use the input side of a removed opto to drive your test circuit the input is 5vdc (or portion of ) to drive the opto, that drives the triac.

REMEMBER, YOU ARE WORKING WITH LEATHAL VOLTAGES...... CAREFUL !!!!

Edited November 11, 2016 by a31ford
Spelling

[DrWizard]

I am planning to run an experiment soon...

I am going to put a high-powered high-voltage (400V 25A) bridge rectifier before the power input to the controller.  The purpose of which is to make all of my half-wave LED strings work at full-wave and be twice as bright.

Since the DC will NOT be filtered, I believe the triacs will still work, and will cut off as they should when the voltage goes to 0v  120 times per second.  I also think the zero crossing detector circuit should continue to work for the same reason.

Where I anticipate a problem is the transformer that drops the 120VAC to a lower voltage to operate the controller.  I will need to isolate that from the power to the triacs, and provide a separate AC power source for it.

[k6ccc]

Like Tricky, that seems like a lot of work for very little benefit.  I think you would be far easier to figure out the rectification in your light strings and convert the strings to full wave.  Depending on how the string is built, it may be fairly easy (may not too).

I would not count on the Triacs shutting off as they may not really get to zero volts.  All it would take is the slightest capacitance (think two wires close to each other) to filter it just enough.

You may find that the transformer that feeds low voltage works just fine because it will still see a fluctuating (semi-AC) on the primary.  Because of the really non sine wave shape it may be less efficient resulting in more heat production.  The power requirement for the control logic is pretty small and has a regulator so that part MAY not need any modification.

 

[a31ford]

Confused... triacs only work on a.c..  and the optocoupler that comes with the board is NOT zero crossing... 

The power transformer on the board will ONLY work with AC.

Technically a triac will work in a dc circuit HOWEVER only one side of the bidirectional SCR pair will be used THEREFOR the triac will need to be de-rated to1/2 of its rated output.

Edited March 22, 2018 by a31ford

[a31ford]

The half wave strings will burn like bacon.... Canac ???? Here boy. Here boy..... ?

[DrWizard]

On 3/22/2018 at 1:29 PM, k6ccc said:

that seems like a lot of work for very little benefit.  I think you would be far easier to figure out the rectification in your light strings and convert the strings to full wave.

I disagree.  I know how to convert a half wave string to full wave, and it's not that hard.  But by rectifying the power before it goes into the controller, I effectively convert 16 strings all at once.  It will be even easier to cut a short extension cord and put a rectifier in the middle of it than clipping the LED string in four spots and inserting 4 separate diodes.

On 3/22/2018 at 1:30 PM, a31ford said:

triacs only work on a.c..  and the optocoupler that comes with the board is NOT zero crossing... 

The power transformer on the board will ONLY work with AC.

Triacs, once turned on, stay on as long there is voltage to the anode, even after removing the signal from the gate.  They turn off only when the anode voltage goes to zero.  So they will work on non-filtered rectified DC.  Likewise, the zero crossing detector doesn't detect the voltage crossing zero, per se, it detects that it has gone to zero (or more accurately, when it is not at zero) so that should continue to work as well.

On 3/22/2018 at 1:29 PM, k6ccc said:

You may find that the transformer that feeds low voltage works just fine because it will still see a fluctuating (semi-AC) on the primary.

Using an AC transformer on full-wave rectified DC results in doubling the frequency (no big deal) and substantially drops the output voltage, and can cause it to overheat.  I will definitely need to provide a separate non-rectified source for the transformer.  OR, better yet, I could just take it out completely and feed the logic circuit with the 5V DC power supply I already have which drives my pixels, arduinos, and other stuff.

On 3/22/2018 at 1:29 PM, k6ccc said:

I would not count on the Triacs shutting off as they may not really get to zero volts.  All it would take is the slightest capacitance (think two wires close to each other) to filter it just enough.

Very good point!!  I had not considered that one!  If I keep the controller close to the lights with short wires, that should not be a problem, but I wasn't planning on doing that.  I keep the majority of my controllers centrally located in the garage.  This actually may be the one factor that keeps my experiment from working. 

On 3/22/2018 at 1:33 PM, a31ford said:

The half wave strings will burn like bacon

Possibly.  It does double their heat output.  If the strings were designed to take advantage of being half wave, and supply peak current a little above what they are rated for (as is oft done with LED matrices) then that could be a problem.  I have converted a few cheap Walmart strings to full wave though,  and they worked fine.  Also, in my sequences, I try not to leave the strings on at 100% for an extended period of time.  If I plan to leave the string on for a while, I tend to do it at 75%.  I especially try to do that on blue, green, and purple strings to reduce failures of the individual LEDs.  I could also make a custom dimming curve to reduce power, but that negates the point of converting to DC to make 'em twice as bright.  Truthfully, I feel everything is plenty as bright as it is, but I thought this would make an interesting experiment.

 

Thanks to everyone for their input!! 

[Mega Arch]

1 hour ago, DrWizard said:

I disagree.  I know how to convert a half wave string to full wave, and it's not that hard.  But by rectifying the power before it goes into the controller, I effectively convert 16 strings all at once.  It will be even easier to cut a short extension cord and put a rectifier in the middle of it than clipping the LED string in four spots and inserting 4 separate diodes.

Triacs, once turned on, stay on as long there is voltage to the anode, even after removing the signal from the gate.  They turn off only when the anode voltage goes to zero.  So they will work on non-filtered rectified DC.  Likewise, the zero crossing detector doesn't detect the voltage crossing zero, per se, it detects that it has gone to zero (or more accurately, when it is not at zero) so that should continue to work as well.

Using an AC transformer on full-wave rectified DC results in doubling the frequency (no big deal) and substantially drops the output voltage, and can cause it to overheat.  I will definitely need to provide a separate non-rectified source for the transformer.  OR, better yet, I could just take it out completely and feed the logic circuit with the 5V DC power supply I already have which drives my pixels, arduinos, and other stuff.

Very good point!!  I had not considered that one!  If I keep the controller close to the lights with short wires, that should not be a problem, but I wasn't planning on doing that.  I keep the majority of my controllers centrally located in the garage.  This actually may be the one factor that keeps my experiment from working. 

Possibly.  It does double their heat output.  If the strings were designed to take advantage of being half wave, and supply peak current a little above what they are rated for (as is oft done with LED matrices) then that could be a problem.  I have converted a few cheap Walmart strings to full wave though,  and they worked fine.  Also, in my sequences, I try not to leave the strings on at 100% for an extended period of time.  If I plan to leave the string on for a while, I tend to do it at 75%.  I especially try to do that on blue, green, and purple strings to reduce failures of the individual LEDs.  I could also make a custom dimming curve to reduce power, but that negates the point of converting to DC to make 'em twice as bright.  Truthfully, I feel everything is plenty as bright as it is, but I thought this would make an interesting experiment.

 

Thanks to everyone for their input!! 

Good grief a lotta work. People on Ebay will buy your existing setup for more than you originally paid. Then take the new found fortune and buy what you need. Seems time would be better spent sequencing or fishing or drinking or something.