LEDS kept blowing out, New solution, Bulb Condoms

[ke7trp]

I have attempted to keep LEDs in some of my old tube gear as replacements for the 44 or 47 bulbs.  I have tried 3mm, 5mm, super brights and the "drake" replacement bayonet style sold on Fleabay.  ALL of them burn out after a month or so.  I have wired them in series and P, Low volts to high volts.  Resistor, diodes in line ect.. Nothing seems to work. After about a month, I look up and realize, 1 or 2 are dead.  I have a feeling its the heat inside the radio cabinet.  I have one LED running on a walwart at 15.5 volts which is above the 12volt rating and its been on for 4 months straight. So it must be the heat or RF killing them off.

I replaced the LEDs with correct bulbs again and then found that Pinball machines use Bulb condoms for the color change. My main reason for these is so the rigs match the rest of the gear in the shack.  At .35 cents each, Why not.

I ordered from this guy:

http://www.pinballlife.com/index.php?p=product&id=311&parent=75

I got the thin and thick. I decided that I like the thin version the best. They have more light output.  They slip right over the bulbs and really work great.  The SP 600 is now red Later, I will switch it back blue to match the audio rack and transmitter.

I orded an assortment of colors and the bill was $10 shipped.  Cheap fun to make your rig the color you want or to make it match other rigs.

C

[Pete, WA2CWA]

I remember Lafayette Radio used something similar on lamps in several of their hi-fi receivers. After several years, some got brittle and cracked, others shrunk around the glass and adhered themselves to the glass. I went to a craft/hobby shop years ago and bought a small stain glass paint kit. Had 5 or 6 colors, 1 or 2 oz (?) bottles with a brush applicator. Cost was about $9 or $10. Worked very well for coloring small lamps to provide new color lighting. Stuff didn't chip or flake off.

[WA1GFZ]

I think you are running way too much current through the leds to blow them out. I have some panel LEDs still working after 30 years.

[ke7trp]

The last batch had 6.7 volts DC on them. They are rated up to 12 volts.  They worked for a month and dead. They last forever if i use them in a SS radio like an Icom. They seem to last forever on the walwart or power supply at various voltages. But if I put them in this SP600, they are gone in short order.  Any suggestions on a circiut to help this?

C

[ke7trp]

These are silicone. They claim they last for many years. But at 35cents. I wont even try to change them if a bulb dies. Just toss it out.

Here is a picture of the SP600 with Red bulbs with the bulb caps.  In person. The red light looks much better and cleaner. The camera makes it look kinda crappy. 

C

[WA1GFZ]

I hope you had a series diode in each string. Feeding AC into them may be a problem. I assume you have LEDS with built in series resistors if rated for 12 volts.
I would use a series diode and an electrolytic cap as a small filter to avoid rectifier noise. Add a .001 uf cap across the diode if you still get noise.
I would use an additional series resistors if the LEDS are in a hot area.
I suspect this isn't your problem in a normal open chassis.

[ke7trp]

Ran them without diode and with the diode.  All these diodes come with build in resistor.  Did not use an electrolitic but did try a bypass cap. Still blew out after a week or a month. I guess some more work is needed to really figure out why. I had a stable 6.7 volts there DC with no ripple detected. 

C

[DMOD]

If it's not the heat killing them and you are running them on pure DC, you may either be experiencing a voltage surge, a negative voltage spike, or RF is getting into them.

I would make my own LED string and calculate the current limit resistor for say 15 milliamps.

Most discreet LED's have a 2.1 volt drop across them and can take up to 25 milliamps. (The exceptions are some of the purple, white, and Blue LEDS which have a 3.1 to 4.0 volts drop across them).

So let's say you are running them off 12Volt DC from your relay supply and every relay has snubber diodes to kill the reverse spike.

For three common LED's in a series string, and running them at 15 milliamps, the series current limit resistor should be:

Rcl = (12 - 3*2.1)/15*10-^3 = 380 ohms. A 330 ohm resistor in series with a 51 resistor should just about do it!

Phil - AC0OB

[WA1GFZ]

The only way I have ever blown a LED is reverse voltage or high current so maybe Phil has a point there could be a transient. Turning off a transformer primary voltage generates a transient. An electrolytic could absorb it.

[N4LTA]

Likely reverse voltage - they usually are specified for Zero reverse voltage - using AC and a dropping resistor won't work for long.

[ke7trp]

I think its the high voltage mute circiut. When you put the radio in send, It cuts off the HV.  I have this on a relay so the receiver will mute.  But when you switch it, The LEDs blink slightly. The Bulbs do not show this at all.  Maybe its to fast for the bulb to report it. 

C

[w1vtp]

My 2 cents: use a DC current source rather than a voltage source to light the leds.  That will help with any surges you may be experiencing.  

Nice look.  Maybe I should try it with my SP-600

Al

[ke7trp]

The red got irritating.  I changed it to green. This worked great as I have the AAde Frequency meter hooked to the Sp600. Now the radio matches the meter.  Its almost exactly the same green when I used the thick bulb caps.

Also pictured is the Gates external modulation meter that TFO built me.  This one got blue LEDs and the gates meter in the rack also got blue caps. 

These have been on for nearly a day now and they show no signs of heating up. They are made of high temp silicone. I think they will last forever and not burn like the old plastic kind.

By the way, I have noticed a very large diffrence in light output between brand new bulbs.  I have a pack of GE's and some other types all from the 50s and 60s plus some new 44s and 47s.  Testing brand new bulbs shows that some are very very bright and some are kinda dim.  I guess its low QC on GE's part or maybe the age?  I sat down and tested a handfull to get a matched set.  The difference would be as much as 50% on some bulbs.

As for running the LEDs on a dedicated source, That is not an option for me. I dont want to modify these old recievers to much. The idea was to put the LEDS in old 44 bases or buy plug in LEDS. This way, nothing is modified or hacked up in the RXer.

C

[k4kyv]

I did a web search for pilot lamps and came up with Ken's Electronics in Michigan.  They list a large variety of types in stock, a much better selection than from Mouser.  Like about every other useful long time "standard" electronic component, regular screw-in and bayonet pilot lamps are getting hard to find and clerks at local electronic parts stores increasingly tend to respond with a blank stare when you ask for one.

Wonder if anyone here  has done business with Ken's and if their service was satisfactory. I notice on their web site they cater heavily to the CB market, but not much mention of amateur radio.

[WBear2GCR]

There are plug in pilot light bulbs made from LEDs that replace incandescants.

Also, I would suggest that a 3pin regulator set up as a current regulator (easy to do, just a resistor) will keep the current in check. Back to back zeners of suitable voltage will kill spikes on the supply line too...

The other option is a single IC that is purpose built for running LEDs - the benefit is that they are often "dimmable"... and they run multiple LEDs.

The only "mod" to the radio would be to intercept the wire(s) that run the bulbs and put the little PCB under the chassis...

Should avoid any bulb blow outs completely.

                       _-_-bear

[ke7trp]

You did not read the post bear.  I was using "plug in" LED bulbs.  These have a three pin regulator and a resistor and STILL burn out in about a month.  They are all over ebay. I even tried the ones sold to Drake people. Those lasted 2 weeks and dead.

I am positive that the issue here is two fold.  Junk Chinese LEDS and RF or a voltage spike.  I have tried just about everything and have gone back to bulbs.

Don, Kens is great.  I have used them for switches, mic cartridges and other hard to find parts for years.

C

[W1DAN]

Hi:

Some good ideas here.

One thing to consider is that standard LED's are current devices. If you wish to use just a naked LED, set up a test rig. Clip lead from your lamp socket to a diode, resistor and then the LED through a test current meter. By increasing the series resistance, find where the current and illumination just starts to drop, then install a  fixed resistor of that value. Running below rated current  should prolong the life of the LED. Also a carefully selected zener diode and current limiting resistor at the supply may be installed to clamp any transients.

Just some thoughts.

GL!

Dan

[k4kyv]

The stock "plate on" and "filaments on" indicator lamps in my Gates BC1-T used 220-volt miniature incandescent lamps.  I  converted the control system to run off 110v DC so the relays don't chatter, and use that same DC voltage source to run the lamps.  I don't have any 220v bulbs, but have  found 110v ones to glow normally but the  have a much shorter life when run on DC.  There is a specific term (which I forget) for this phenomenon, but the filament tends to evaporate in little rings, less evenly and less uniformly when DC is applied, eventually causing a weak spot in the filament and ultimately, burn-out.

I picked up a bunch of white, high output LEDs at Dayton last year, and plan "one of these days" to construct a substitute on top of the base of a defunct bulb, with appropriate current limiting resistor.  Hopefully this will last longer than the incandescent bulbs.

I may also try using 220v bulbs if I can find some.  Even on DC, they should last a long time running on half voltage, assuming the light output is sufficient for the purpose.

I use nothing but 12-14v bulbs in my 75A-4.  They seem to last for ever.  I have had some in for over a decade and they are still going strong.  I use a #53 to replace the little short round bulbs, and a 4-digit number to replace the #44/47s. Not only do the bulbs last much longer; I prefer their warm amber glow to the harsh bright white light from stock 6.3v lamps.

[WA1GFZ]

Don, Why don't you use neon bulbs in the high voltage application

[k4kyv]

Probably wouldn't be bright enough.

BTW, does anyone know if a NE-51, a small neon lamp with a bayonet base and glass envelope identical to those of the #47 incandescent pilot light bulb, has a built-in current limiting resistor?

[Pete, WA2CWA]

Probably wouldn't be bright enough.

BTW, does anyone know if a NE-51, a small neon lamp with a bayonet base and glass envelope identical to those of the #47 incandescent pilot light bulb, has a built-in current limiting resistor?

http://www.donsbulbs.com/cgi-bin/r/b.pl/ne51~usa.html >>SEE Left Side of page

[WA1GFZ]

I don't think it has a limiting resistor

[WBear2GCR]

Whoa... it wasn't clear to me that you were using the commercial LED replacement bulbs. Yeah, sounds like RF or else a BIG spike. Heat is an outside possibility...

Ok, so... that would get us a transzorb across and a small shunt cap for RF at each socket... maybe a ferrite bead too ??

               _-_-bear

[ke7trp]

I just wonder if its the heat.  The commerical replacements have a resistor and a three leg transistor inside. I ripped a couple apart. The Transistor is dead. The LED lit up with 3 volts applied directly to it.  So the failure here is that tranistor.
C

[DMOD]

I suspect the three-pin device is the TO-92 or TO-92L packaged version of the '05 or '08 regulator.

http://pdf1.alldatasheet.com/datasheet-pdf/view/238082/TAITRON/LM78LM08H-XX-AR.html

You must be exceeding the power rating of that device.

Do you have a part number for that LED bulb replacement?