Driving a led with mains. / General Science and Electronics / Forums

Forums

4hv.org :: Forums :: General Science and Electronics

« Previous topic | Next topic »

Driving a led with mains.

first previous 2 3 4 5 next last

Move Thread

LAN_403

Sulaiman

Wed Sept 18 2013, 06:17PM

Registered Member #162 Joined: Mon Feb 13 2006, 10:25AM
Location: United Kingdom
Posts: 3141

Experimentonomen, you're only partially correct there,
the Chinese/eBay led lamps with a small number of leds use an smpsu
the large cob/corn led lamps have diodes and resistors to power long series led strings.

Ash Small

Wed Sept 18 2013, 06:30PM

Registered Member #3414 Joined: Sun Nov 14 2010, 05:05PM
Location: UK
Posts: 4245

Dr. Slack wrote ...

Ash, have you actually looked at Bjorn's link? Very thorough, informative, the thread should have stopped with that.

I'm going to choose not to accuse you of trolling, yet. If C2 is 1000x C1, then the voltage across C2 will be 0.24v rms (non US) or half that (US), not enough to light a LED. Use the provided information, and think.

I have 50 LEDs in series, being driven at 20mA 160v from a series C from 240v AC into a bridge rectifier with a 120uF 330v photoflash cap across it, as I've drawn earlier in the thread, lighting my windowless hallway. Quite efficient, quite Christmassy. Runs cool, no transformer, so not at all stupid.

Yes, I have read Bjorn's link.

The point I'm making (or the question I'm asking) is about capacitance voltage dividers.

OK, I didn't check the forward voltage drop across the LED, which is around two volts.

OK, the numbers were wrong, but if C2 is 100 times C1 then the voltage across it is ~2.4V, which is sufficient to power most LED's.

I've not done the maths required to determine what value C2 needs to be to hold sufficient charge to power an LED for 0.01 seconds (half of one cycle @50Hz), but COULD this circuit be built using just a fuse (albeit a small one), two resistors, two capacitors and two LED's?

EDIT: I guess I'll start with ~2mF for C2 (1 Farad=1Amp@1Volt for 1Second (from memory), 0.01Sx200mA=2mF. Seems like a good place to start).

EDIT: I made another mistake, current through the LED is ~20mA, not 200mA (I thought it didn't look right), so a good starting point for C2 would be ~200uF, and C1 will be ~2uF.

johnf

Thu Sept 19 2013, 09:58AM

Registered Member #230 Joined: Tue Feb 21 2006, 08:01PM
Location: Gracefield lower Hutt
Posts: 284

Ok I should have been a little more specific in my previous thread I meant two LED strings back to back equalling most of the applied voltage.
This is how the super cheap Mains LED bulbs work ie 110 white LEDs how ever I see from the failed unit in front of me there is a 22 ohm brick red resistor in series (presumably as a fuse as these resistors are cheaper than fuses.). However there is also a resistor across the capacitor of 1 Meg I thought to presumably to discharge the cap when removing the bulb from the fixture. Cap is 1uF 400volt ie 105n.
this bulb run for 1 year then half the leds stopped and very quickly afterwards the other chain stopped (a few days).

Bulb Chinese rated at 8 watts 240 volts

Ash Small

Fri Sept 20 2013, 09:22AM

Registered Member #3414 Joined: Sun Nov 14 2010, 05:05PM
Location: UK
Posts: 4245

Well, I've got to here now:

1379668951 3414 FT157237 Led Mains Indicator

The 1Mohm bleeder resistor passes ~1/4mA, the 25ohm resistor should pass~20mA when the 200uF capacitor is charged to 1/2V above the LED forward voltage drop, and the 5ohm resistor is 'a few ohms', as reccomended by Dr Slack.

It obviously requires a fuse as well, but I'll build it and see what happens

(I'm expecting to have to adjust some values, though, in order to achieve optimum operation. This is just a starting point.)

BigBad

Fri Sept 20 2013, 07:33PM

Registered Member #2529 Joined: Thu Dec 10 2009, 02:43AM
Location:
Posts: 600

Why don't you LTSpice it?

Ash Small

Fri Sept 20 2013, 08:40PM

Registered Member #3414 Joined: Sun Nov 14 2010, 05:05PM
Location: UK
Posts: 4245

BigBad wrote ...

Why don't you LTSpice it?

I did consider it, but that would probably take me far longer to get the simulation right than just wire it up and see if it goes bang.

If anyone who is more familiar with LTSpice than I am could do a quick sim, I'd appreciate it.

I've tried to err on the side of safety, as far as guestimating values, and I'm sure if I play around with the values I can get it more efficient, but I'm pretty certain the values are 'in the ballpark'.

I've ordered a 2uF motor run capacitor (250V) and two 100uF non-polar electrolytics (100V), which I'll pick up when I go into town on Sunday. I'll pick up some fuses too. The worst that can happen is exploding capacitors and blown fuses.

EDIT If anyone does do a sim, I'd suggest leaving out the bleeder resistor initially.

Ash Small

Sat Sept 21 2013, 01:02PM

Registered Member #3414 Joined: Sun Nov 14 2010, 05:05PM
Location: UK
Posts: 4245

Sorry about the DP, but some new information:

I had some time to kill this morning, so started playing with LTSpice. I've not yet been able to set the forward voltage drop of the diodes, but I'm getting the following result:

The waveform shows the current through R2 in green, but doesn't take into account the forward voltage drop of the diodes, but it does look as if the circuit works ok.

The blue trace is current through R1.

How do I set the voltage drop, anyone?

EDIT: Looks like it 'assumes' the voltage drop across the diodes is 0.8V

Looks like I've got around it, I selected a Fairchild QTLP690C, and changed R2 to 1 ohm, and obtained this waveform:

I think I changed R1 to 1 ohm as well, but I'll run some sims later to optimize the value of R1.

EDIT EDIT: Well, I played around with it some more, removed both resistors, and set C1 at 1uF, and C2 at 10uF, and obtained the trace below. The LED's don't need surge protection, as they are not in series with the capacitors, so, if the capacitors can withstand any mains surges if switched on mid-cycle, no resistors are required, or so I hope to prove.

The voltages and currents through both LED's are well within limits, and power through the LED's is, from what I can make out, around 0.01W or maybe less, although there are bound to be other losses.

Looks like I've ordered the wrong capacitors now

Question, Are these PP audio capacitors, rated for 630V going to be up to the job?

Dr. Slack

Sun Sept 22 2013, 06:47AM

Registered Member #72 Joined: Thu Feb 09 2006, 08:29AM
Location: UK St. Albans
Posts: 1659

Question, Are these PP audio capacitors, rated for 630V going to be up to the job?

After years of bodging around, I've finally hit on a really time-saving modus operandum. Rather than ask the question 'with this underrated thing do', I now start off with 'how do the professionals do it? Are there standards that exist?' OK, if the latter is too expensive, I can then think again, but it's a real time and equipment saver.

Check out X and Y rated capacitors. They are designed and approved for direct connections to the mains, including recovering from and failing safe under extreme transient conditions. It's up to you to find what the difference is, and choose the appropriate type.

Ash Small

Sun Sept 22 2013, 06:40PM

Registered Member #3414 Joined: Sun Nov 14 2010, 05:05PM
Location: UK
Posts: 4245

That's a fair enough answer. I picked up one X rated capacitor today, along with the others
.
The two audiophile capacitors I picked up today are PP MKP. I'm going to try connecting them to the mains and see what happens. They are both rated for 630V (DC), so should handle 50Hz mains voltage (240V), it's just a question of whether they can handle the current surge if switched on mid-cycle. If they do blow, I'll get some motor run capacitors, or lighting capacitors, which are designed to be connected directly to the mains.

The reason I asked yesterday is that Maplin don't provide manufacturer's part numbers, so I couldn't look up the data sheet. I'll do that this evening, and then edit this post.

EDIT: The capacitors are LCR MKP 10 WF X

The closest I can find online are LCR MKP 10 WE AX, which are the same dimensions. The pulse rating of these is 1x 10 MÎ© - uF @ Rated Voltage, although the part number isn't identical. They must have significant inductance, so should provide some impedance to surges. Hopefully I'll get them wired up tomorrow.

]mkp-10-ax-wf.pdf[/file]

Mattski

Sun Sept 22 2013, 07:40PM

Registered Member #1792 Joined: Fri Oct 31 2008, 08:12PM
Location: University of California
Posts: 527

wrote ...

EDIT: Looks like it 'assumes' the voltage drop across the diodes is 0.8V

Ash, LTSpice uses Spice models for components like diodes. You can insert a spice directive such as '.model Dcustom D (Is=0.1p Rs=1 n=1)' to create the model 'Dcustom', then set 'Dcustom' as the value of the diode instead of 'D'. This uses a model I=Is*(exp(qV/nkT)-1) plus a series resistance.

Also if you right click on the diode in LTSpice you can select "pick new diode" and select from several models for commercial diodes that are built in.

first previous 2 3 4 5 next last

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob