20 mA LED SMPS buggy

Jun 11, 2012 31 Replies

I always try to avoid those nasty shocks, but occasionally I get one. Probably because I'm not as careful as when trying to make a jacobs ladder with my MOT :).

I consider using a MOSFET but atm I use bipolars because IMHO they're more simple to drive from the TL494 output.

I get that value, too, with my equations. However, I only have a sawtooth up to a certain point. Further increases in duty cycle ... see my OP and

formatting link
formatting link

That's the voltage across a 1 ohm sense resistor in series with the flyback primary.

That's not a matter of base charge, IMO.

But maybe this problem is current related and won't show up once I switch to 325V, 100-200 kHz and 8 mA peak current. Maybe I should just try. Then I will have to deal with turn on/off time.

I get a relatively sharp decrease which levels at about 300 mA during the *on time*.

I tried:

12V - 470 ohm - C2 - E2 - base - 100 ohm - ground

same problem.

With ...

12V - 1K + base - emitter - ground | C2 - E2 - ground

... the transistor is off when Q2 is on, so the duty cycle is reversed. With the pot all the way down I get from the minimum dead time a minimum transistor on time. The LEDs get about 1 mA (estimate from the brightness) and slowly fade out completely - without modifying the duty cycle. Meanwhile the average primary current rises to 130 mA when the LEDs are out and keeps rising until I disconnect power when the transistor get's hot to the touch. I haven't been to the lab to watch this on the scope.

Bernhard

"Ian Field" wrote in news:sqJBr.42046$ snipped-for-privacy@fx16.am:

dealextreme has LED "supplies" designed to work from 240VAC,perhaps he could tinker with one to change it's current output from amps to what he needs for his low power LEDs. or modify a Li-ion battery charger to supply the low mA he needs for his project;they also use a transformerless line-to-DUT supply,and are fairly inexpensive.

at minimum,they would give him the IC type and circuit needed to do what he wants.

Jim Yanik jyanik at localnet dot com

Have you tried a capacitive voltage divider and connect the bridge, surge protection resistor and LED chain to the midpoint of the divider.

Of course, there should be quite a few LEDs in the chain (10-20) to keep the parallel capacitor at a reasonable size.

I chain of big LEDs (0.5-3 A) could be driven through an choke with reasonable inductances, e.g. through fluorescent ballasts (at least in the 230 V world).

There was recently a selection in the news & new products section of various LED drive modules, one in particular was a module with the form factor of the ballast for a linear florescent - it made big of being potted & having a high IP rating.

Potting seems to be a growing trend just lately.

"Ian Field" wrote in news:Uc0Cr.67235$ snipped-for-privacy@fx29.am:

My $7 dealextreme Li-ion 18650 2-cell charger is not potted. it's also very lightweight,so there's no XFMR in it(I haven't gone into it...) and the chip and circuit board can probably be accessed and modded. (unless it's chip-on-board!)

it runs off 100-240VAC.

Jim Yanik jyanik at localnet dot com

There is almost without a doubt a HF transformer in there, just not a mains-frequency transformer.

If it's this one:

formatting link

.. it's indicated to be Class II (double-insulated).

By the look of the adverts; planar transformers are starting to make an impact on the marketplace.

Back in the 80's Sony did a CTV PSU that switched a hefty electrolytic between the bridge rectifier & HT rails, IIRC it used GTO thyristors - very nasty things to faultfind.

"Ian Field" wrote in news:2%1Cr.89391$ snipped-for-privacy@fx24.am:

I opened it up,it definitely contains a small HF XFMR,about an inch square,0.5" high. 2 5-terminal SMD ICs labeled "5465",a TL431,and several xstrs. a pretty lousy soldering job,too.

Jim Yanik jyanik at localnet dot com

Is it 10uS/division? If it is, it is slow:

That wont work because C2, E2 is not a negation of C1, E1 - sorry.

Glenn

...

...

The turn off seems to be 2uS?

If it is 10uS/div, you turn the transistor on for 60uS?

Are you sure that you have phased the transformer for flyback?:

formatting link

Glenn

put a large low ESR capacitor in parallell with the LEDS

it's effectivly a capacitive divider, so if you're dropping 240V to 7V make the parallel cap 240/7 times the size of the series cap. then the series resistor only needs to be sized to protect the rectifier

?? 100% natural --- Posted via news://freenews.netfront.net/ - Complaints to news@netfront.net

An AC Xc potential divider might work - using a bridge rectifier and damping the rail with an electrolyitic might also help a little - but there is a serious warning! - If the LED chain is intermittent, even for a split second and the cap will develop enough voltage to punch out the chain of LEDs - intermittent for more than a split second could burst the cap.

Join the Discussion

Have something to add? Share your thoughts — no account required.

Didn't find your answer?

Ask the community — no account required