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 5 6 7 8 next last

Move Thread

LAN_403

Ash Small

Wed Sept 25 2013, 09:24AM

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

I will attempt to measure the inductance of the capacitors later today.

The pulse rating of the closest spec capacitors I've found is 1x 10 MÎ© - uF @ Rated
Voltage, which I assume translates as 10Mohm per uF, at the rated voltage, so the 10uF one has a pulse rating of 100Mohm. I'm guessing this means it's inductance is greater than a nano-Henry.

It is just a roll of ~30mm wide foil and PP. I think it may well be longer than 10 feet, though

BigBad

Wed Sept 25 2013, 09:29AM

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

I think if how many nanohenries it is actually matters for your circuit there's something very wrong!!!

All I'm saying is you shouldn't use estimates that are out by several orders of magnitude.

But by all means add an inductor to smooth out the current through the LED.

Incidentally, in LTSpice you can add inductance to a capacitor by setting the properties, right clicking on it.

Ash Small

Wed Sept 25 2013, 10:28AM

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

BigBad wrote ...

I think if how many nanohenries it is actually matters for your circuit there's something very wrong!!!

All I'm saying is you shouldn't use estimates that are out by several orders of magnitude.

But by all means add an inductor to smooth out the current through the LED.

Incidentally, in LTSpice you can add inductance to a capacitor by setting the properties, right clicking on it.

The inductance of the capacitors only affects the surge current when the circuit is switched on at any voltage not equat to zero.

The value only affects the frequency and amplitude of the spikes/ringing.

I tried values from the nH to mH range, all had similar effects.

If it blocks ringing at the frequency depicted, it should block the higher frequencies associated with a lower value inductance.

I agree that I may be a couple of orders of magnitude out, but it wasn't necessary to know the frequency of the ringing in order to eliminate it, all that was required was to isolate the LED from the ringing completely.

I do agree that I need to measure the inductance, and then enter 'realistic' values before the design is complete, though. I also think I can significantly reduce, or even eliminate completely, the two remaining resistors on the output side, so that the only losses in the circuit are from losses in the capacitors and diodes (and resistance of the wire), as stated in the OP's 'terms of reference'.

Steve Conner

Wed Sept 25 2013, 11:39AM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

That MOhm-uF thing has nothing to do with the pulse rating of the capacitor. It's a measure of the DC leakage current, how long the capacitor would take to discharge just sitting there.

The pulse rating is normally given in volts per microsecond, which amounts to the same thing as a peak current rating in amps. This has nothing to do with inductance either, it's a matter of how much thermal shock and Lorentz force the thin metal film inside the capacitor can stand before it explodes.

If you assume about 10nH for the ESL, that should do it.

Ash Small

Wed Sept 25 2013, 12:12PM

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

Fair enough, Steve, I'll take your word for it, but this is how it appears on the spec sheet:

Ash Small

Wed Sept 25 2013, 12:47PM

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

New information:

Here are the waveforms across the LED with 1nH/uF.

Steve Conner

Wed Sept 25 2013, 01:00PM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

The spec sheet is wrong, end of story.

Ash Small

Wed Sept 25 2013, 03:12PM

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

Actually, some ringing is still getting to the LED during the first 50 or so micro-seconds after turning on mid-phase:

Ash Small

Wed Sept 25 2013, 07:01PM

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

New information:

First measurements seem to be suggesting @70nH for the 10uF capacitor, maybe higher.

EDIT: Around 100nF. around 160kHz is where the amplitude 'bottoms out' on the 'scope, which gives ~100nH if the capacitor is 10uF.

EDIT: when I connect the two capacitors in series, it suggests my previous measurement was too low. I'm getting ~350kHz, which equates to a total series inductance of over 200nH.

EDIT: After a bit more 'twiddling', I't appears that the inductance of the 10uF cap is over 100nH, The Cornell Dubilier datasheet here:

list a 1uF, 600V capacitor with identical dimensions as mine having an inductance of 25nH....These appear to be 'realistic' values for this type of capacitor.

I'll re-do everything tomorrow, and see if I can get closer to a definite value.

Edit: I'm thinking that a 'combined' inductance value of ~133nH is 'in the ballpark'.

Ash Small

Thu Sept 26 2013, 10:54AM

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

Here's another simulation with some fairly realistic figures for self-inductance of the capacitors.

The ringing lasts longer, and has a different amplitude and frequency, but doesn't present any problems as far as I can see.

Maybe I can measure the inductances more accurately later, but I think it's reasonable to assume that, whatever the actual value, it won't present any problems (assuming the simulation is correct).

Next job is to try to eliminate the resistors. I'll update this post by editing as and when I make any progress.

EDIT: well, that was pretty easy, I removed the resistors, and reduced C1 from 1uF to 0.5uF, and got these results:

The ringing now takes a completely different form, but it doesn't look like it will present any problems.

I'll try some different values for capacitor self-inductance and see if it makes much difference.

EDIT: Here's another screenshot showing how different 'realistic' capacitor self-inductance values affect the spikes/ringing.

Here's the schematic as it stands at the moment:

1380235763 3414 FT1630 Led Mains Indicator

last updated 25th september 2013

EDIT: This circuit may need bigger diodes than the 1N4148's, in order to cope with the surge when switching on mid-cycle.

MURS320 diodes seem fine, but are slower. It seems any 15, 25 or 50 nS diode rated for a few amps will do it.

first previous 5 6 7 8 next last

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