Another crazy noise-making device!

[adibrook]

You may remember a while back i tried to build a flourescent light amp, as described in some article somewhere, by wrapping a coil around a flourescent light bulb and hoping that putting audio through the coil would modulate the current passing thorugh the flourescent lamp.

Well...it didnt work at all. All i got were wierd noises. However, i have a friend who likes wierd noises (he manages a record label that produces starnge electronic music amongst other things). Hes allways bugging me for new electronic devices which make noises.

So, i got back to investigating this phenomenon. It seems that the arc in a flourescent bulb is actually pretty noisy.

All the other circuits i built to investigate this phenomenon before just had different flourescent lamps connected to a 250VDC power supply through a big valve output transformer. They all made funny noises but nothgin special.

However, an important event happened when i noticed the noises changed when i powered down the device. As the anode voltage gets less the noise changes. So, the sound is voltage-dependant, which could be VERY cool.

So, i decided to test exactly how voltage dependant the sound is, by building this:

http://img.photobucket.com/albums/v4...elampsynth.gif

Hers how it works:

The bottom bit is a normal amp. I allredy had that built on my prototyping board, which is why non of the components are marked. I cant remember what they are...but the amp DOES definetly work.

Next is a DC power supply powered by a variac so i can control the voltage. the rectifier and caps are all rated 500VDC or there abouts so they are well within spec.

The top filiment of the flourescent lamp is blown by the way (although in the schematic i drew it complete for the sake of...erm...symetry?). So, the two filiment supports make the main anode and a starting cathode. The voltage from the amps heaters is rectified to power the starting relay, which when you press a button momenterely makes the ''starting cathode'' negative which ionises the gas in the lamp and the lamp lights. Only a quick tap on the starting switch is nesesery to start the lamp.

Now comes the interasting bit. It seems that the device produces very wierd noises when the lamp is at a lower voltage that it should be. At about 80-120v on the anode the coolest stuff happens. You get everythign from allmost white noise to sine waves, and lots of very wierd effects when different harmonics mix etc (like a high pitched sine wave seemingly modulated by a much lower pitch sine wave).

No doubt the flourescent lamp will be just as active in the RF spectrum (or maybe not) but i surrounded it with an earthed metal shield just to be sure, and the wires to it are short and shielded so it doesnt act like a transmitter. Doesnt seem to screw up my radio's reception.

This is only a prototype. It can do with several improvements, such as a nicer starting system that heats the cathode filiment momenterily (longer tube life) and an electronic anode voltage control instead of a variac (biiig power valve?).

By the way...this has no other use exept a novelty for making strange noises. But...its very good at that. My friend is totally blown away by it, saying its crazier than any synth setup hes ever herd, and wants a ''real'' non prototype version. But before building one i want to investigate this more.

So...any opinions anyone?

[Panrock]

Love it! What a fabulous idea!

Steve

[GMB]

I would suggest that you add a resistor to control the current because I think the tube current will massively escalate if you turn the variac up too much and it will just burn everything out. I hope the strange sounds are not that of the tube being destroyed!

My guess is that the tube and the transformer is making an oscillator but that it is operating in a chaotic mode. Chaotic systems will oscillate at various frequencies which can be made to appear quite regular, and as the chaos increases become more like noise even though what it is doing is not actually random.

[adibrook]

Well, the tube isnt exactly being run under ideal conditions. So far we burnt out two tubes. Well, not exactly burnt them out, just the noises they produce arnt as cool, and the arc is harder to start. I broke one of the tubes to see, and the ''anode'' melted completly, and metallised its surroudnign area.

Next time i will put in a 1ohm resistor and mesure the voltage across it in the anode circuit to see how much current exactly its drawing.

I dont think the variac is too stressed, since it remains totally cool, and i want to replace the valve o/p transformer with some mains transofrmer or something, just to make sure i dont kill a valve transformer by accidnet. The choke (which is the primery from a mains starnsformer from some modern device) is also nice and cool. And the bridge rectifier module also doesnt look stressed.

The flourescent tube DOES look stressed, it runs hotter than usual on soem settings (altough at low voltage its not hot at all). But to be honest...flourescent tubes are cheap and avalible. These are only the little 4 or 6w types which are avalible for £1.99 from wilkinson. One tube lasts a few hours. With refinments to the circuit it may be extended to a maybe 10 or 20 hrs, which is good for me.

I want to have a way of statting the arc without blowing one of the filiments. Then both the cathode and the anode will be intact filiments, which should raise tube life and preformance alot.

I know that to start it you need to heat up the cathode filiment momenterily. But does anyone know what voltage/current they need?

Thanks.

[Mersey_Swimmer]

Firstly, you really do need to ballast your tube. Like any other gas-discharge device a fluorescent tube has a negative resistance characteristic and should you not limit the current? Well back in 1965 as an inquisitive teenager I fired up a tube then shorted out the ballast. BANG! Tube shattered before my bench fuse went.

You can use a small mains light bulb to ballast a DC circuit, for your experiment a 230v 25W would probably suffice. Just wire in series with the DC supply to the tube. A ballast lamp won't make starting more difficult as it'll only drop voltage once the tube has struck.

[G8DLG]

The heating current should not be more than the the tube operating current.Ie for a 2ft or 4ft tube its 400ma. This may be difficult to achieve with switch start circuit,since you dont have the voltage drop across the tube to limit the current.Ideally the supply voltage shoud be at least twice the tube volts,a ckoke is also essential to give a high voltage kick when the starter switch opens,(as with AC).Using a fluorescent on DC mains requires a suitable resistor as ballast. The problem with DC operation is that the tube will go black on one end after some hours,so you will have to reverse the tube now and then
Efficiency is a little higher on DC,light output is not pulsating.

[evingar]

I did have a very brief "tinker around" with the original idea of using a florescent tube as an amplifier element - Basically found what Aidy found - any effect is masked by a load of "hash".

I did wonder about using one of those miniature neon lamps Must give it a go when I get a minute.

[adibrook]

I think the idea of the original flourescent tube amp is its push/pull, using two flourescent tubes, and supposedly the noise from them would cancel it self out like that. But...i dunno. I think the noise in the tubes may not be identicle, and therefore the output will still be noisy since to cancel out both parts have to be identicle and opposite.

I will experiment with different ballast circuits and see how it effects the operation of the device.

However, i also want to develop some kind of non-variac voltage control method.

Woudl this just be a case of putting a BIIG power valve between the flourescent tube and its HT supply, and then putting a variable voltage on the valves control grid?

[Mersey_Swimmer]

Originally Posted by adibrook

However, i also want to develop some kind of non-variac voltage control method.

Woudl this just be a case of putting a BIIG power valve between the flourescent tube and its HT supply, and then putting a variable voltage on the valves control grid?

Best stick to the Variac, it's by far the safest and most efficient method. Using a valve, big rheostat or anything else to regulate the DC side will result in wasteful power dissipation as heat - lots of it.

Solid-state control of the AC side is also a no-no as should your triac or SCR go short.........! And the wideband noise that results from "chopping up" the AC mains would be intolerable in your experiment.

[daviddeakin]

Originally Posted by adibrook

Woudl this just be a case of putting a BIIG power valve between the flourescent tube and its HT supply, and then putting a variable voltage on the valves control grid?

That is more or less the idea behind a series voltage regulator circuit yes.