replacement neon (or LED) indicators

[aptsys]

For a few mA, we need about 100kΩ.
Xc = 1/(2πfC), so 330nF. I'd chuck in a resistor in series to be a fuse. If you use 2 leds back to back in parallel, the reverse bias voltage is kept low, and no flicker.

You need the resistor in series to limit the inrush otherwise the LED won't last more than a few switch cycles. The inrush current is limited only by the LED otherwise!

 

[JohnW2]

You need the resistor in series to limit the inrush otherwise the LED won't last more than a few switch cycles. The inrush current is limited only by the LED otherwise!

That's got me thinking. For the fist few microseconds, the capacitor is going to do essentially nothing to limit current (particularly in the 'worst-case' scenario of switching on at the moment of peak supply voltage), so if one wanted to limit the inrush current to a figure which was not dramatically higher than the LED's operating current, one would need a fairly substantial resistor, probably a significant proportion of the value which would be required if there were no capacitor. However, the capacitor value capacitive reactance would then have to be appreciably smaller (than if there were no resistor) in order to achieve the desired running current. In turn, that would mean that the 'running' current limitation would be substantially down to the resistor, taking us back at least partially to the 'power consumption' (and resistor heating) issue.

I suppose it might be possible to find a usable compromise which didn't result in too high an initial inrush but which also was fairly 'efficient' whilst running - but a lot would depend on how tolerant (in terms of life expectancy) the LEDs were to very high but very brief inrush currents.

Kind Regards, John
Edit: typo corrected

 

[JohnW2]

Another problem with neons is they need a stimulus to strike. They are very reluctant to strike if in total darkness, They need a few photons to trigger ionisation in the gas.

Interesting. I've never heard of, or thought about, that one.

Kind Regards, John

 

[JohnD]

I suppose it might be possible to find a usable compromise which didn't result in too high an initial inrush but which also was fairly 'efficient' whilst running - but a lot would depend on how tolerant (in terms of life expectancy) the LEDs were to very high but very brief inrush currents.

the circuitry in LED lightbulbs is very basic, and I hear the lower the wattage, the cooler it runs and the longer it will last. So I'd guess that as it gets miniaturised and churned out by the billion, mains-voltage indicator lamps will also get cheaper and smaller. The idea of soldering individual components into every lamp is ludicrous.

 

[JohnW2]

the circuitry in LED lightbulbs is very basic, and I hear the lower the wattage, the cooler it runs and the longer it will last. So I'd guess that as it gets miniaturised and churned out by the billion, mains-voltage indicator lamps will also get cheaper and smaller. The idea of soldering individual components into every lamp is ludicrous.

Agreed, and to some extent what you say is undoubtedly true, but there are some stumbling blocks in terms of laws of physics which can never be changed ...

... so long as we stay with LED technology, I can't see that we will ever get away from the fact that the LED element wants, say, 3-5V (most commonly nearer to 3V), whereas the supply is around 325V peak.

Something therefore has to be done to limit current. As aptys has implied, if that is done primarily with a capacitor, then there will be an unacceptable inrush current. On the other hand, if it is done primarily with a resistor, then the amount of power dissipated in that resistor will be high (better part of 1W with a 3mA LED), not just a waste of energy but probably enough to melt an LED encapsulation. As I said, there may be an acceptable 'compromise', but I'm not convinced that what could be achieved would necessarily be a very good compromise.

If one could, physically and cost-effectively, get an "SMPSU" within an LED package, that would probably do it - but I'm certainly not holding my breath for that!

Kind Regards, John

 

[winston1]

That's got me thinking. For the fist few microseconds, the capacitor is going to do essentially nothing to limit current (particularly in the 'worst-case' scenario of switching on at the moment of peak supply voltage), so if one wanted to limit the inrush current to a figure which was not dramatically higher than the LED's operating current, one would need a fairly substantial resistor, probably a significant proportion of the value which would be required if there were no capacitor. However, the capacitor value would then have to be appreciably smaller (than if there were no resistor) in order to achieve the desired running current. In turn, that would mean that the 'running' current limitation would be substantially down to the resistor, taking us back at least partially to the 'power consumption' (and resistor heating) issue.

Do a bit more thinking about that.

If you add a resistor the capacitive reactance would have to be smaller. So the capacitor value would have to be LARGER.

 

[JohnW2]

Do a bit more thinking about that. If you add a resistor the capacitive reactance would have to be smaller. So the capacitor value would have to be LARGER.

Yep, a sloppy typo error on my part When my fingers typed "capacitor value", my brain had instructed them to type "capacitive reactance".

However, once that typo is corrected (which I've done), nothing else I wrote is changed.

Kind Regards, John

 

[FrodoOne]

I have a couple of these

which are working fine, but the neons have gone dim. To save the expense of new ones (they are a non-standard size), is it possible to get replacement mains voltage neons? I have an idea they have a resistor in the connecting wire to drop the voltage.

I realise that there have been many responses to this question and I am sorry if this has been covered by anyone else but the answer as to why these "neon" indicators dim or "fail" is the quality of the resistors limiting the current through them.

The resistors concerned need to be in the range of up to Megohms!
These are difficult to make for the few cents/pence usually expected.
Unfortunately, manufacturers tend to use the relatively cheap "carbon granule" resistors, which initially have the correct resistance value but they depend on "point contacts" between the individual granules to provide this resistance. In operation, some of these point contacts tend to go "open circuit", which results in a higher value of resistance and a cascading effect, resulting in the open circuiting of more point contacts - and so on.

The solution is to use better quality (metal film) "dropping resistors".

This brings to mind the situation when we had TVs with cathode ray tubes, which required both vertical and horizontal deflection.

My memory is now a bit rusty but some vertical deflection circuits in these TVs involved one or two quite high value resistive components.
When these cheap resistors crept "high", the vertical deflection would shrink.
When replacing them, apart from using the best quality that I could obtain, if space was available I often replaced one of these with four in series parallel, to reduce the stress on each component - which probably gave the TV more than four times its expected life, as determined by that problem.

 

[JohnW2]

I realise that there have been many responses to this question and I am sorry if this has been covered by anyone else but the answer as to why these "neon" indicators dim or "fail" is the quality of the resistors limiting the current through them. .... The resistors concerned need to be in the range of up to Megohms! ... These are difficult to make for the few cents/pence usually expected. ... The solution is to use better quality (metal film) "dropping resistors"

I'm not at all sure about that. The resistors are typically about 220kΩ, sometimes up to about 600kΩ, and even cheap carbon resistors of such values generally keep their value for decades in service.

I've just played with three such neon indicators, taken out of service because of dim or dead neons. In all three cases I replaced the resistors (220kΩ in two cases, 270kΩ in the third) with brand new components and that did nothing to improve the performance of any of the neons - and all three of the 'old' resistors still had measured resistances close to their nominal values.

Kind Regards, John

 

[333rocky333]

Asked in Wf , I didnt have numbers with me and none of the part numbers from there older mk book were on the system, sorry.
I will do more enquires

 

[JohnD]

thanks very much!

But I have managed to get a few neons with resistors. They're smaller than the originals, but I'll give them a try when I'm redecorating the kitchen.

 

[Echo the husky]

Interesting. I've never heard of, or thought about, that one.

Kind Regards, John

It's a strange effect, I was showing it to an apprentice a few weeks back, it is much more noticeable on an ageing neon, I switched the room light off and it flickered on only occasionally but with the room light on it lit all the time.

 

[JohnW2]

It's a strange effect, I was showing it to an apprentice a few weeks back, it is much more noticeable on an ageing neon, I switched the room light off and it flickered on only occasionally but with the room light on it lit all the time.

Interesting!

Merry Christmas!

Kind Regards, John

 

[SimonH2]

Can one get a pair of back-to-back LEDs in a single encapsulation?

Yes, you can get a variety of multi-chip devices in single packages - including back-to-back pairs

For the fist few microseconds, the capacitor is going to do essentially nothing to limit current ...

Bear in mind that LEDs have multiple current ratings - one of which is a pulse power rating which is significantly higher than the steady-state DC current rating. So an inrush current might not be too much of a problem - you'd need to study the specs and do some calculations.