Series vs parallel LEDs in photoswept devices

[Mark Hammer]

There are some cases where an LDR-based phaser or whatnot involves a single light source, whether incandescent bulb or LED, and an array of LDRs around it.  And there are other cases where an LFO drives several LEDs in series, each of which illuminates its own LDR.

When the LEDs are placed in series like that, doesn't each LED receive a different amount of current, and light up just a little differently, or am I misunderstanding things?  I was just wondering whether it would serve any useful purpose to have parallel buffered outputs driving the multiple LEDs in parallel instead of having the LEDs in series.  Maybe it is just the case that the slight difference between LEDs in series pales in comparison behind the variation in LDRs?

[puretube]

series-LEDs add their Uf, which leaves less "headroom" for the LFO to sweep...

[Mark Hammer]

So does that "round off" the LFO sweep in some way?

[puretube]

not at all: it cuts off a part of the wave...

[Mark Hammer]

Like a clipped waveform or some other sort of headroom limitation?

I'm assuming that the waveform alteration is only for one half-cycle, such that the sweep is asymmetrical.

[puretube]

imagine an LFO sinewave from 0V to 8V:
the wave will start lighting up the LEDs
from the point where it`s value is higher than the sum of the Uf`s;
and the LEDs will fall dark as soon as the momentary LFO-curve
drops below that summed up Uf.

Of course you can give the LFO a DC threshold to "ride on" (say a +4V "baseline"),
but then the sinewave amplitude (=variation in brightness) has to be reduced to
4V max  (from +4V to +8V), if you want it "clean".

[Mark Hammer]

Sadly, I think you've lost me there.  "Uf" means capacitance, right, not some electronic parameter I'm unfamiliar with?  Or is that forward voltage?

If I understand what you're saying, the combined forward voltages of the LEDs in series (just like diodes in series in a clipper circuit) prevent the LFO generator's output from translating into light until the LFO amplitude reaches the voltage equal to the combined forward voltage.  At that point, the LEDs conduct, like any diode would, and they light up for whatever is left of the rest of the LFO waveform.

Seems to me that would produce a rather choppy sweep.  Or is the purpose to give the LDRs a chance to "recover in darkness", making use of that recovery curve to simulate a pure sinusoidal "turnaround" at that end of the sweep?

[Nasse]

If I understood it the voltage is divided by the number of leds, but of course same current goes trough them... but I sort of think it might be coarser control that way because the active thing that the often crude lfo gives is just voltage... I need sleep

But I believe all superbright leds are not equal linearity, new types coming and going

[puretube]

Uf=forward voltage

[Mark Hammer]

Okay.  Then I think I understood.

What still has not been answered is my original question.  Is there something better about series LEDsn than parallel ones?

If the advantage they might provide is because of asymmetrical sweep, then that suggests that the sweep of a single LED lighting 2 or 4 LDRS might be "improved" by placing several diodes in series with the LED to change the turn-on point?

[davebungo]

Uf=forward voltage

Why didn't you just say Vf?

[puretube]

Uf=forward voltage

Why didn't you just say Vf?

coz Ufk = filament-to-cathode voltage...
(cathode = Kathode in German, btw; the old language barrier)

for me: Ub=9V means : the battery voltage is 9Volt(s).
           

[davebungo]

Okay.  Then I think I understood.

What still has not been answered is my original question.  Is there something better about series LEDsn than parallel ones?

If the advantage they might provide is because of asymmetrical sweep, then that suggests that the sweep of a single LED lighting 2 or 4 LDRS might be "improved" by placing several diodes in series with the LED to change the turn-on point?

As far as the LEDs are concerned, they don't know the difference.  i.e. as long as they pass the same current (say Iled) they will output the same amount of light.  But here is the rub:  If you connect them in parallel, you need 2 * Iled mA in total.  If you put them in series, you only need 1 * Iled mA but of course you need a voltage high enough to drive 2 Vfs.  So really, it all depends on your drive circuit.  It's nothing to do with the LEDs.  

The only other thing I can think of is that if they are in series they will always both pass the same amount of current, whereas it could differ slightly if they are in parallel.  This may be an advantage if you have different LED types in the circuit.  i.e. ones which requires different amounts of drive.

[johngreene]

Okay.  Then I think I understood.

What still has not been answered is my original question.  Is there something better about series LEDsn than parallel ones?

If the advantage they might provide is because of asymmetrical sweep, then that suggests that the sweep of a single LED lighting 2 or 4 LDRS might be "improved" by placing several diodes in series with the LED to change the turn-on point?

I think the short answer is:

In series, all the LEDs receive the same amount of current, actually share the same current.  The only variance in light will be due to the difference in the light generating characterisics of the individual diodes. The Vf shouldn't be a problem if you are driving them properly, i.e. with a current source. The key point here is all are driven by a single source.

In parallel, you have to 'balance' them. Either with series resistors or have them each driven by a separate current source. So, you would use 4 times the current for 4 LEDS in parallel than you would if they were connected in series.

--john

[davebungo]

Uf=forward voltage

Why didn't you just say Vf?

coz Ufk = filament-to-cathode voltage...
(cathode = Kathode in German, btw; the old language barrier)

for me: Ub=9V means : the battery voltage is 9Volt(s).
           

I remember using "Ub" in my apprentice days when I was at college come to think of it.  But still, now I would always use V especially in a forum where some may not remember or know about the "old" nomenclature.

[Mark Hammer]

Thanks you gentlemen.  Much much clearer now.  The sense I am getting is that LED illumination:
a) may well be more consistent from LED to LED (and certainly no LESS consistent) when they are used in series compared to parallel,
b) consumes less current when in series than when in parallel

Both very good reasons why it's what I tend to see in posted schems.

[davebungo]

I found this which is interesting (one of thousands of similar application notes probably):
http://www.maxim-ic.com/appnotes.cfm/appnote_number/1804/

[toneman]

Good link there Dave!!!   

[petemoore]

  Why is it we want all the LED's to light up equally and synchroniously?
  Would it be a bad thing if resistances sweeping phase stages began and ended their sweeps at separate intervals?

[Mark Hammer]

Depends.  Remember that the number of degrees of phase shift are summed across stages.  Achieving a notch requires that the number of degrees of shift at frequency F add up to 180 (or multiples).  One assumes that having the same amount of phase shift in each stage at frequency F increases the likelihood of being able to generate a notch.  Of course, variation in LDRs doesn't mean that precision sweep of one's LEDs will produce perfectly matched phase shift response across all stages.

Matching of FETs in phase shifters is done so that all FETs used are able to change in response to the LFO, and that none have reached the endpoint of their sweep before the others have.  That generally achieves a more desirable and consistent sweep across the entire range, compared to having 3 stages make it to the end zone and one stage crapping out at the 20 yard line.  Although it may well be nigh impossible to have flawless stage-to-stage matching of phase shift when using discrete LED/LDR-based sweep, you still want to have a smooth sweep that feels continuous, so having the LEDs turn on and off at the same time is probably a good idea.  That doesn't mean that disparate LED illumination is pointless.  It may well do something interesting, but I suspect that "ideal" phaser sweep really wants to have truly synchronized LEDs.