I have an old piece of test equipment that had a sealed neon bulb for an ON indicator. This pilot light is dead. I cant seem to find a replacement that will fit, which means I will have to drill out the panel to fit a replacement. I would like to keep the panel looking original, and I also dont want to go thru the hassle of drilling it, which would require front panel removal, to avoid damaging other components while drilling.
What I want to do, is just mount a LED behind the original light. In other words, cut away the old neon bulb and place a LED right behind the bezel of the old pilot light.
This all seems simple enough, and I know that using a LED requires a resistor in series with the LED. However, I am attaching this to the power line (120VAC).
Correct me if I'm wrong, but it seems to me that a LED requires DC. Does this mean I need to install a power diode (such as a 1N007) in series with the LED and resistor? Or does it also need a filter capacitor?
these indicators usually contain an NE2 or similar neon globe, if you can open it you could replace that.
It's probably easier to replace the neon with a new neon, but yeah.
the capacitor is not needed (but if fitted will reduce flicker) resistor and diode are needed. a bridge rectifier around the LED instead of the single rectifier will give more brightness for the same heat (and also reduce flicker)
When I tried casting out nines I made a hash of it.
A red/green one with a suitable diode resistor bypass pair in series could even mimic the orange glow of a neon. The circuit shown above though will blow the poor LED to pieces the first time a serious fast HV glitch comes down the line.
The OP might want to look for old scrap gear using neons though. New spares of most commmon sorts are still available in the UK :
They are in the US too but you need to buy 20 at a time and pay less:
The differentiator action of the RC + reverse diode clamps the voltage across the LED and time-limits the current spike in the first circuit. Yeah the current can jump very high for a short period of time in response to a several kilovolt spike but the LED is not going to fail from taking an amp for a few tens of microseconds. Reverse breakdown is the bigger danger
Silicon diodes have microsecond peak amperage ratings some two orders of magnitude greater than their maximum average current ratings. I suspect with LEDS (typically with Vfs some 3 or 4 times higher, this will be more like a factor of 25 than a 100.
Luckily, in this application, we won't be pushing our indicator LEDs anywhere near their maximum average current limits so we can get away with specifying a current limit resistor value corresponding to 10% of the impedance of the dropper capacitor chosen to set the current to circa one or two milliamp through our 20 to 50 mA rated red LED from the 70s and 80s.
Around a decade ago, I wired up three strategically placed 13A mains sockets wired up to my SmartUPS2000 (located in our basement) and wanted to provide an indicator lamp on each (two doubles and, afaicr, a single - ICBA to check what I think is a single right now) not only to indicate their powered up state in the event of a power loss but also to mark them out as "Protected Supply" sockets. I decided against using neon lamps with 220K resistor since not only were they a little bit 'naff', they also suffered a limited lifetime, hence the more unusual at that time small 3mm dia red LEDs.
These red LEDs had been recovered from obsoleted industry grade modem circuit boards that I'd purchased either in the late 70s or mid 80s (the LED chips were mounted on a ceramic header, topped off with a bead of red translucent epoxy - a 'quality' of construction that's not been seen since the 80s). Suffice to say, these weren't the most efficient choice of LED even back in the 90s, let alone the noughties.
However, for this use, they were efficient enough even using a simple series capacitor of 33nF and a 1 or 10 K surge limiting resistor (ICBA to open one up to check what resistor value I actually chose ten years or so back) with an anti-parallel diode to protect the LED.
Although this requires higher C & lower R values than a full wave rectifier circuit and despite the use of such low efficiency LEDs as these vintage ones, even I considered this extra complexity a step too far so single anti-parallel diode it was.
In this circuit, even a 50v PIV rated 1N4001 would have sufficed - Hell! another vintage red LED would have done! I had plenty to spare and space for its slightly greater bulk than that of a 1N400x series diode - I feel, in hindsight, that I missed a trick in not saving those three
1N4000 series diodes for better things. :-(
Anyhow, these cobbled up LED indicators are still going strong after a decade's worth of continuous service. When it came to my choice of surge limiting resistor, my only concern was the maximum conceivable worst case switch on surge condition of a +350v charged capacitor as a result of being switched off at just the worst possible moment, applying the sum total of 700v by being switched on at the most conceivably extreme case of a -350v peak in the mains supply cycle.
I felt that such a conceivably worst case scenario would nicely cover the even rarer microsecond wide kilovolt transient events that might occur in real life so didn't make any further allowances than that. After some ten years of trouble free service, it seems my choice of anti-surge resistance value was entirely justified.
With modern red or green LEDs (more so if you can use an anti-parallel red/green LED), you can probably use something like a 33K 4n7F ballast with a cheap 1n4001 anti-parallel diode circuit to replace the neon lamp without fear of destruction by high voltage transients on a 240vac supply (use 15K and 10nF on a 120vac supply). If you use a blue (or 'white') LED, you can eliminate the C and multiply the R values by 50 to 100. :-)