Extra LED dial lamps

[David 51]

Hi Everyone
Wanted to put some extra dial lights on the Radio I'm restoring as the two 6.3v AC 300ma lamps fitted don't light up the FM scale which is below the others.
The lamps fitted are AC in parallel with the heater chain. I wanted to fit an LED with a 390 ohm resistor in inverse parallel to each of the lamps & have them underneath the bulbs to light up the FM scale. So it would be easy to take back to original if desired; would this cause any problems?

The radio is a Nordmende Fidelio 593D
Thanks
David

[TrevorG3VLF]

Looks reasonable.

I have not done this, others can perhaps help more than me.
The LED current will depend on the LED voltage (not specified). The LED will pass current for half the time so the current could be a bit higher than being run from DC.
A disc ceramic capacitor across the LEDs may be advisable in case of interference.

[Herald1360]

You will need two LEDs in inverse parallel in series with a resistor across each lamp. The resistor will depend on what LED current you need and what sort of LED.... white ones have rather more forward voltage than coloured ones.

I think the reverse voltage rating of LEDs is higher than their forward voltage, if not some additional diodes in series with each LED will fix it.

The extra 100mA or so loading on the heater circuit is unlikely to upset anything.

[TrevorG3VLF]

The LEDs should be in inverse parallel with the resistor in series with this.

[Herald1360]

Originally Posted by Herald1360

You will need two LEDs in inverse parallel in series with a resistor across each lamp.

Originally Posted by TrevorG3VLF

The LEDs should be in inverse parallel with the resistor in series with this.

Maybe I should have added some brackets somewhere.

[McMurdo]

AFAIK its more to do with the fact that a reverse biased LED is passing no current so the dropper resistor of course presents the full voltage to the LED. LEDs only have a low reverse voltage capability. If you havent got room for a second LED in reverse, a small 1N4148 etc would do in inverse parallel. Depending on your eyes it might flicker...though the one in my doorbell push doesn't.

[julie_m]

But a reverse-biased LED is passing current, or at least it will if you apply enough voltage to it; because it does a fair impression of a Zener diode. Rectifier operation breaks down somewhere between 4 and 6 Volts, depending upon LED chemistry. If you are going to drive a single LED from AC and rely on its self-rectification property, you need to choose the ballast resistor so as to limit the power dissipation in reverse breakdown.

As the Zener voltage is less than the forward voltage, when two LEDs are wired back-to-back, only one is conducting; the other one is still blocking.

[Herald1360]

But the idea is that there are two LEDs in inverse parallel. Provided the reverse breakdown of each is higher than the forward drop of the other then everything's hunky dory with one LED conducting on +ve half cycles and the other on -ve ones. If not, stick a 1N4148 or similar in same direction series with each LED to do the reverse blocking bit.

Which is simply an expanded version of what I originally suggested in post #3.

[majoconz]

Provided the reverse breakdown of each is higher than the forward drop of the other

But IMHO it's not. I'm with Herald on this one. I tried the same trick in a Sony CFR 5090 where the lamps had blown. The LED's didn't last long. The LEDs need to be inverse parallel, but with their own diode in series. The reverse breakdown is pitiful on the bright white LEDs I used, I measured a good one at Vf = 3.2v and Vr = 1.6v - they were from a local supplier who said "Spec's? Who knows" when I asked about Vr!

[broadgage]

Presuming a very low voltage supply like a 6.3 volt heater circuit, an LED may be simply connected connected in series with a suitable dropper resistance, when the supply polarity is "correct" it will light, and on the opposite half cycle it will be subjected to reverse voltage, which matters not at such a low voltage.

It would however be better to use 2 LEDS connected in inverese parralel with each other, and the pair in series with a suitable resistance. If the 2 LEDS light the same area, then flicker is reduced.

You could use two pairs of LEDs in series with each other, thereby halving the total current used as compared to two pairs in paralel.

I doubt that the extra current from the heater winding is important.

[julie_m]

Well, if the reverse breakdown voltage is lower than the forward voltage drop, then the power dissipation in the reverse breakdown state ought to be lower than in forward bias (though, note: with more voltage appearing across the ballast resistor, the current through the ensemble will increase). So, as long as there was sufficient voltage available, you could just connect two such LEDs in (inverse) series for AC operation.

[David 51]

Hi
Tried it using 290 ohm resistors in series with the LED's one across each dial lamp inverse to each outher & it works, no flicker, both LED's have simular light output to the filament bulbs & are warm colour, so it looks good, havn't noticed any bad effects like interferance. If they were going to blow would they go open circuit or short circuit.
Thanks
David

[G8HQP Dave]

Unplanned reverse breakdown can sometimes damage a semiconductor, so best not to use this routinely except for zeners and other devices which are known to cope OK. Heat is not the only possible damage mechanism.

I didn't know that reverse breakdown is at a lower voltage than forward conduction for (some?) white LEDs. We are always learning something new on here! For 'normal' red/green etc. LEDs the reverse parallel, plus series resistor, connection should be fine.

[Herald1360]

A cunning plan if it works and a bean counter's delight.

A quick check on a few RS stock through hole white LEDs suggest Vr of 5V and Vf around 3.8V so inverse parallel should be OK.


Since there's a resistor in series, if they were to blow open, one LED would stay on (for a while?) blow short, both go out. Either way unless the resistor is seriously underrated, no smoke will be let out

[AlanC]

Could you also have fed the LED's from a bridge recitifier; which would surely make all the worries about Vr go away? You could connect it in series with the chain and it would preserve the AC continuity, or so one would think...
Cheers,
Alan

[julie_m]

Originally Posted by G8HQP Dave

Unplanned reverse breakdown can sometimes damage a semiconductor, so best not to use this routinely except for zeners and other devices which are known to cope OK.

It's always been my experience that LEDs, B-E junctions and similar have recovered fine from such (mis-)treatment where I have kept the power dissipation down. (And I've even seen it done in commercial devices.)

So have I just been lucky, then?

[Herald1360]

Originally Posted by AlanC

Could you also have fed the LED's from a bridge recitifier

But that's 4 diodes per LED pair instead of at worst, 2.

[G8HQP Dave]

Originally Posted by ajs_derby

It's always been my experience that LEDs, B-E junctions and similar have recovered fine from such (mis-)treatment where I have kept the power dissipation down. (And I've even seen it done in commercial devices.)

So have I just been lucky, then?

I did say 'sometimes' and 'except . . for those known to be OK'. For example, reverse breakdown can increase noise in some semiconductors, while not causing catastrophic failure. Others survive with no harm done.

[Billy T]

I just wire the LEDs in inverse-parallel and feed them both with 6.3 vac via a single series resistor. I set the resistor value for the level of brightness required, and since this can't be calculated and depends on the installation, I start with a relatively high resistance value and reduce it in arbitrary steps until I get the desired result. Since the current requirements are low, a 'resistor wheel' allows speedy selection of the required value, starting with a high value and reducing until the required brightness is achieved.

Since one LED is always in forward conduction, the reverse voltage across the non-conducting LED cannot exceed the forward voltage of the conducting LED. If more than two LEDs are required, a separate circuit is necessary to avoid having LEDs of the same polarity connected in parallel. For lack (as yet) of any cogent reason to do so, I have not used capacitors in parallel with the LEDs, so the jury is still out on that refinement.

If 6.3 vac was not available, I'd use whatever ac voltage was, or connect the leds in series and steal a few mA from the HT, adjusting the series resistor value accordingly in each case.

In several hundred hours of operation I've had no problems, so if there is a flaw in my reasoning, the evidence is yet to show its face.

Cheers

Billy

[kalee20]

It's fine to connect two LED's in inverse parallel, as Billy and others have done, for AC operation. Only snag is, you have to put up with the flicker - obviously you can't put a capacitor across the pair! If the LED's are close together, the overall flicker frequency is doubled.

Me, the last time I wanted to run a LED from AC, I used a single series diode, a series resistor, and a capacitor across the LED. But that was 25 years ago. Since then, I've either used incandescent bulbs from AC, or I've had a DC source already existing for LED's.

As this is a radio application, why not put the LED's in the DC supply anyway? You could connect as part of the output valve cathode bias perhaps, thus getting power 'free'. Or connect in series with the HT rail.

Granted that the LED's wouldn't light until the radio's warmed up, however from the first post, you already have a filament lamp so you can see that you have power immediately you switch on, so there should not be a problem. And if the LED's ever light unnaturally bright, or dim out, you know there is something wrong!