EM34 LED A new approach

[Christoffrad]

EM34 LED

Followers of this forum may be aware of the inventive design by Colin Boggis in 2008 for a solid state approach to replace the EM34 Tuning Indicator. His solution used a diode array and LM3914 to mimic the EM34 display.
My design provides another solution aimed at providing a display that may be regarded as giving a better appearance whether the radio is on or off because the target cone and central button of the valve are retained. However it must be made clear that this solution does not provide any signal strength indication. It is purely a cosmetic effect.
The problem of the diode array is that the display surface and central button are lost and the colour may not be true. This design uses a single diode of the correct wavelength. This is located inside the mechanics of the original valve. The diode is located behind the centre black button to provide illumination on the original valve target cone in order to provide an appearance that looks correct under all situations even though it will not provide the moving vanes effect.

This approach assumes that a defective valve is available.
It is prepared as follows. The valve glass envelope must be opened in order to reveal the inner workings. I was fortunate to be able to call on a local company with a diamond cutter however is you are not able to recover the original glass I suggest the use of a 25mm diameter plastic test tube cut to length (obtainable on-line).
The next task is to carefully remove the centre heater/ cathode assembly using suitable cutters and pliers. Try to leave access to the heater connecting wires near the base of the valve as these can be used to pick up the 6.3v ac supply to power the new circuit. Although I removed the entire glass from the base, this is probably unnecessary if care is taken to ensure that all electrodes do not contact the new circuit. Finally, carefully enlarge the centre hole in the little mica support just behind the target cone. It will also be necessary to slightly bend open the cone where the heater went through. Bend each quadrant just enough to allow a 5mm LED to penetrate. The hole in the mica should be just enough so that the shoulders of the LED sit on the underside of the mica.

The new construction: I have established that in order to achieve a perfect colour match you need an LED that has a green wavelength of 520nm. I obtained one via a well known auction website described as; Super Bright 5mm Green LED LL1503PUSG1-301 520nm 12000mcd 30 deg
This diode would normally be run at 20mA but I found that for accurate illumination I should run it at less than 0.5mA ! I also found that in order to get good dispersion of light onto the original target cone I needed to reflect the light. I achieved this using double sided sticky foam onto which aluminium foil is attached to provide a mirror like surface. I used two thicknesses of Halfords double sided tape. Stick the foil on one side then cut a circular pad that will just fit inside the black button on the original valve. To locate it I suggest you shape a thin piece of wire so that you can stick it to the circular piece of tape in order to hook it under the button and push it firmly into place ( by feeding something through the LED hole to apply pressure). This is quite a delicate operation. You should test the light dispersion by locating the LED in position and feeding it from a 9v supply via a 10k resistor. I decided that an LED with a diffused lens would have been best but as one was not available I cut a small piece of plastic from an everyday milk container and wrapped it around the LED body. When the LED was suitably positioned (LED shoulder sitting on the mica) I applied a little epoxy glue to cement the LED to the mica.

The new circuit: I established the required current to give a comparable display to that of a traditional indicator valve. If you do not use the exact diode that I have you may need to experiment with needed forward current and precise positioning of the diode. I used four low current diodes to form a bridge rectifier followed by a 1k resistor feeding an electrolytic capacitor (10uF is enough) to provide the dc supply. A further 10k series current limiting resistor is used to define the brightness and ensure that during trials you can’t attach the LED directly to a fully charged capacitor…pop!
The circuit components are loaded onto a small piece of vero board. Not the copper strip variety but rather the plain sort so that components are wire interconnected on the back. This allows close packing of the components. To mount the board I noticed that if you remove the getter plate from the valve (by tearing the plate from its mounting bar) you have a rather useful bar that can be bent through 90deg and onto which the board can be lodged and the bar then bent back to hold the board captive. Finally I used epoxy glue to cement the glass back into place. This type of glue has the rigidity to support the fragile glass.
The completed indicator was reassembled into my restored Philips 170a radio on which I have just replaced the cloth and feedback transformer (but that’s another story) all I need now is a Philips badge for it.
Note the last pic shows a comparison of this EM34 LED with a 6U5G indicator.

[paulsherwin]

Nice job. You could perhaps vary the brightness according to the AGC voltage, using an FET or op-amp. In fact you may be able to power the LED directly from the AGC line given the low current required.

[marlowradio]

That is certainly an excellent replication, would it be possible to post the circuit diagram as I would like a go at a couple of duff magic eyes I have

Many Thanks
Steve

[dseymo1]

Cct diagram is included in the OP's pictures.

[marlowradio]

My mistake, I just saw that after sending the request

[Christoffrad]

The circuit diagram is shown in the first 2 photos. I have more photos available.
Anyone got a Philips badge to spare? (Red with blue'Philips')

[ThePillenwerfer]

You can make an ordinary LED defused by rubbing it with fine sand paper.

[paulsherwin]

You can also crack the moulded lens off with a pair of sidecutters.

[Refugee]

I would put the LED in a hole drilled through a bit of scrap plastic and sandpaper the lens off.
That would ensure that the crack does not make shadows.

[G6Tanuki]

Originally Posted by paulsherwin

Nice job. You could perhaps vary the brightness according to the AGC voltage, using an FET or op-amp. In fact you may be able to power the LED directly from the AGC line given the low current required.

Rather than risk loading the AGC line with the LED, you could always wire it in a 'bridge' circuit between the cathode of an AGC-controlled IF amp and the cathode of the output valve.

(the output-valve will have a constant cathode-voltage due to its bias-resistor; as the AGC action occurs the voltage at the cathode of the AGC-controlled IF stage will fall, so increasing the voltage available for the LED)

This was a classic way of operating a S-meter in old communications receivers.

"Tune for maximum brightness", is instinctively logical.

[Christoffrad]

Thanks for that AGC / OP cathode idea. Sound like it's well worth a try.

Chris

[bikelectro]

This seems a really great solution.
A few years ago I bought a boxed NOS EM1 on buy it now or best offer - I offered and got it for about £30!
Its gorgeous but I cannot bear to use it!
If anyone has any duff EM1 or similar, i'd love to do this, and maybe put my good one up for auction! I broke my old one by accident.
They are just too precious and short - lived to leave on in a radio! After only a few hours use you can already see the marks in the phosphor!

[bikelectro]

Well i got an old EM1 and did the job - although not perfect, i am very happy with the result. I can recommend it although it does take 2 - 3 hours!
First i removed the base and broke the pip to let air in, then i wrapped some NiChrome wire around the valve. Passed current through so it glowed red hot and broke open the envelope. Not as neat a split as i’d have liked - but good enough.
I followed most of what was done for the EM34 which is much the same except i flattened the LED and filed it to make it a mat dull surface and painted around the side with black Humbrol to stop light being so visible out the side. Also even though i think my LED was the same, I found i needed it brighter so i ended up with about 400 ohms on the second resistor - the LED has 2.7 volts across it and i think is below its limits ok. I would have liked to leave “service access” but to give it strength i had to epoxy glue it together after a soak test. This is a dreaded side contact base and as only heater pins are used i removed some pins leaving four only - this makes the valve easier to remove from the socket. The colour is just right! I will try to add pics....

[Christoffrad]

I also like your idea to paint the sides of the LED and produce a flat matt end. It seems to work well. Those component values are obviously not critical. It may be beneficial to make the first resistor 470R to have a well charged C (I suppose one could take a look with a scope) then choose the appropriate current limiting resistor to follow. I had it in mind to perhaps try the suggested idea of driving it from the AGC as a brightness tuning indicator and that is why I used the smaller C value this time so that the time constant is shorter. But I probably won't bother!

[bikelectro]

Painting is a good idea - though even in the radio, if you look to the side you can see the top of the LED, but with the rough texture it is not as bright. The light on the target cone is not totally even , but I'm delighted as I have the lovely green glow looking much as it would with the real eye - this is a great solution. You can see I was able to twist the disc to help with gluing a foil disc onto it.
There is someone on the web showing how to open lightbulbs with the Nichrome wire.
I scraped off the red paint first and BE SURE to release the vacumn first and wear safety glasses when cracking the glass open - just in case it should break violently!

[bikelectro]

The AGC idea sounds great, but I think on a slightly weak station that you still may like to listen to - the glow may be a bit dim, and not so pretty?