Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.
If you break one open you will inevitably find two PCBs, one aluminum backed with several LEDs on it and one conventional board with a bridge rectifier and current regulator on it. It is easy to think of them as a single LED but their operation is actually quite complicated, hence the wide voltage range they can operate over.
Unfortunately, pulsing the supply to them doesn't give their internal regulators chance to start up properly.
Now I tried to dim the LED bulbs in series with the resistive dummy load and such way it works perfectly: no flickering and the BUX is cold as well.
The LED bulbs still have this step in current and a little resonance but otherwise everything looks good.
So maybe there is some oscillation when direct driving the "LEDs".
Tomorrow I'll try some snubbers.
I tried some snubbers on the load but it didn't help to make the BUX cooler.
I'll try some other mods with this layout and meanwhile I'll build another
dimmer with a zero cross synced trailing edge kind of operation just to
check whether it helps to eliminate my current issues.
Out of curiousoty I measured the LED bulb current
on the same series resistor but driving them directly
from the mains at the current looks like this.
So I guess it's definitely designed for phase based dimming.
My dimmer works with resistive load perfectly.
I'll try it with a MOSFET as well.
Interesting to see on the current waveform that it starts with a large current and then it settles to a lower value.
And even the ringing is there.
After that I doubt that it's managable to fine dim it in the lower region right after the startup.
But we will see...
There is nothing stopping you using an NE555 and your transistor switch to create a phase control signal. Just wire the '555 in monostable mode and use it's output to switch the transistor. You can reset it with the line waveform and turn the transistor on after an adjustable delay. Essentially doing what a conventional triac phase controller does.
Yes, I'll do exactly this: using a simple 555 with zero-cross triggering + delay to create a trailing edge dimmer for my LED bulbs.
I'll report about the results... ;-)
Wow, I tried a IRF730 MOSFET as switching device and it is much better now...!
It can drive any of my previous loads perfectly without getting even warm.
But being such an ideal and fast switch, now at the lowest PWM level my LED bulbs
are starting to shine at ~40%... The impulse is so short and even so they are quite bright. :-o
So as the startup has a lot of current, the dimming doesn't produce a linear
control of the lightness but follows exactly that current pattern in time.
So I guess the next step is to build this zero-cross synced 555 based trailing edge AC dimmer.
My last chance is that starting the "turn on" on a low voltage may help to smoothly dim the LED bulbs at lower levels.
I attached my design and the predicted waveforms with the lowest and mid level dimming. In spice it works fine.
A question: I used a trick to have a simple zero cross detector which provide a short pulse
but I'am not sure: leading the mains via a 1n capacitor can kill the transistor base-emitter junction
or as it's a small pulse which starts at zero so it won't do any harm?
(I designed a 2nd version of that part using a resistor + zener combo to protect the transistor if it blows in reality... )
Aside from the danger if the capacitor goes short, I can't see how it works at all.
The 1nF capacitor will just charge up, surely you want to feed it from AC to work reliably.
It's simple: the 1nF conducts right after the zero crossing to open the transistor.
Then the transistor pulls down the voltage on its collector.
There is also another 1nF capacitor after that to generate just a pull-down pulse.
And the lefty 1nF will be discharged via the 470k when the (rectified) "AC" is going down to zero again.
If it will work in reality I guess this is a much simpler detector as the usual (with AC optos and so on...).
I just finished the completly new v2.0 of my dimmer and (rarely I have this experience... but )
everything is perfect for the very first time...! :razz:
The dimming is now perfect without any flicker, it goes from 0% to 100%
even the controlling linearity is better, follows the pot nicely.
My biggest surprise was this fully custom designed zero-cross detection method which
works perfectly, I atteched the scope about how it works, I guess it's rather precise... :-D
I attached the final schematic as well which I guess is very simple and cheap.
And some photos about my prototype layout with the PCB (Paper Circuit Board). :-D
Just a little update: there was only 2 issues I'd like to solve but they are ok now.
I attached the new schematic with the 2 changes highlighted.
1) The 2W 100k resistor at the PS was just a starting value and a waste of energy.
I tried higher values and fortunately even a 470k is enough
so now the dissipation is only < 250mW which is "cool"..!
2) Every once and awhile (once in half an hour) there was some "flash" kind
of pulse like effect (the light level jumped on max for a very short time)
which I guess was coming from the mains.
I tried a lot of filtering at the mains, at the LED bulbs, at the MOSFET but nothing helped...
Then I increased the 47k (on the left-top which leads the "AC" to detect
the zero-cross) to 1Meg and now it's perfect.
With these mods now I guess it's working perfectly.
It runs on a very low power level, so there is no parts getting warm at all.
it's likely the 47k was flashing over ( or breaking down in flashes ) as no doubt it is not rated for 325Vpk, the 1Meg will likely last longer - but a 500V rated part would be the engineering choice ...
- - - Updated - - -
Also - local ripple control for hot water cylinders can cause this effect on simple ckts - moving to 1Meg would tend to filter that signal out ...
How do you mean the pot wiring?
It's just a variable resistor in this layout from 0R to 500k.
LEDs: I'll use it just for 5-10W GU10 LED bulbs nothing extra.
But I even tested it with two 50W halogen GU10 bulbs as well.
Output images: due to automatic brightness controll, I guess it wouldn't be so dramatic.
I tried to take a video with my phone but it looks like the light level is almost constant.
What exactly would you like to see?
This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
By continuing to use this site, you are consenting to our use of cookies.