Torch bulbs as modulated light sources

[Panrock]

When I was at school in the 1960s, my very first mechanical "8-line" Nipkow television used a torch bulb as a light source. No, it never worked properly but this was before I'd ever heard of things like equalisation. Light bulbs had the advantage of being relatively bright compared to my little '90v' neon indicators and of course there were no LEDs in those days.

I wonder how fast you could effectively modulate a tiny pea bulb of low thermal capacity by using some DC bias and successively boosting the higher frequencies? What sort of phase delay characteristic would have to be applied to the equalising? I would guess the blue component of the light would modulate faster than the red, and a lot faster than the infra-red (heat).

Maybe it would then be fast enough for an 8-line picture?

Steve O

[AncientBrit]

Hi Steve,

I waited to see if Steve A would pick up on this one as he's more knowledgeable than I but I guess he's busy.

From my recollections about an article back in the 60's for construction of a photophone I think the max upper limit is 1kHz but more realistically I'd say 500Hz.

You could of course set up a little experiment to measure an actual sample.

Connect one end of the bulb to a 6v supply, the other end to the collector of say a 2N3055 NPN transistor
Connect the emitter via 47R to 0v
Feed the base from the wiper of a 1k pot wired across the 6v supply.

This forms a crude constant current generator to feed the bulb.

Feed tone in from a generator via a 47uF cap to the base of the transistor.
Slowly advance the pot from 0v to set some dc through the bulb
Don't blow the bulb!

View the output of a club photocell on a scope and point the bulb at the photocell.
Add some simple optical screening (cardboard tube)

Carry out a "sqeak" using the sine output of the generator
If the signal is clipped reduce the sine amplitude from the generator or increase the dc.
Increase the sine amp at higher frequencies to see how much boost you'd need to maintain a level response

Switch to square wave and see what the risetime is like.

Note that the above assumes that the detector is at least 10x as fast in response as the bulb.
From experience I think that will be the case.

Alternatively use a faster photodiode in place of the dome cell.

Report your findings back here.

As for equalisation circuits that I'll have to leave to others.


Good luck,


Graham

[Panrock]

Thanks Graham.

From memory, putting 2K2 across the 'club' dome cell ensures it has a good response for NBTV in photo-voltaic mode.

I'll have a play with this when next I get a mo. I would think though that smearing on picture will be almost impossible to remove when using a torch-type bulb, because of the thermal lag.

You should be able compensate for slowness of response of the bulb on the leading edge by boosting the hf there - in other words forcing the bulb to heat up faster by hitting it with a big, short sharp pulse. But - by the same token - I can't see how you can force it to cool down faster on the trailing edge. So smearing you will get.

Unless this effect can be masked by some weird form of phase-shifted equalisation? Or forced cooling?

Steve O

[Steve Anderson]

Steve. Graham is correct, it's been hell here dealing with builders and the inevitable errors they make. Why do people look at a drawing but don't actually read it? The same applies I find to e-mails.

Anyway...Chris Long did a lot on this subject many years ago and it's all there on his web-site...but right now it appears to be down...when it's up again I'll post the link.

My only question is why? With LEDs, Luxeons etc. why bother with a bulb? Maybe the novelty factor?

You may find those of higher voltage and low-wattage are better in terms of thermal response, say a 24V/1W bulb as opposed to a 6V version. I will add I don't have any experience in modulating bulbs so I'm only guessing. It could also be quite non-linear as the light (heat) output isn't proportional to the applied voltage.

Steve A.

[M3DVQ]

I would imagine (from a position of total ignorance on the subject you understand!) that with some fancy circuitry you could make the rising edge very sharp by pumping a greater current through the filament for the brief moment between "off" and "on" to heat the filament up more rapidly. However there will be no way to improve the lag when switching from "on" to "off", it will cool in its own sweet time

[AncientBrit]

@ Steve O.

I just spotted a silly mistake in my simple cct I recommended to you.

There should be a standoff resistor, say 4k7 from pot slider to transistor base otherwise the pot will load the tone generator at either end of its travel.

The cap is still directly connected to the transistor base from the generator.

Cheers,

Graham

[Peter S Spooner]

When I was at school in the 1960s, my very first mechanical "8-line" Nipkow television used a torch bulb as a light source. No, it never worked properly but this was before I'd ever heard of things like equalisation. flashlight had the advantage of being relatively bright compared to my little '90v' neon indicators and of course there were no LEDs in those days.

I wonder how fast you could effectively modulate a tiny pea bulb of low thermal capacity by using some DC bias and successively boosting the higher frequencies? What sort of phase delay characteristic would have to be applied to the equalising? I would guess the blue component of the light would modulate faster than the red, and a lot faster than the infra-red (heat).

Maybe it would then be fast enough for an 8-line picture?

Steve O

Link one end of the light to a 6v provide, the other end to the enthusiast of say a 2N3055 NPN transistor
Connect the emitter via 47R to 0v
Feed the platform from the windsheild wiper of a 1k pot wired across the 6v provide.

[Harry Dalek]

Link one end of the light to a 6v provide, the other end to the enthusiast of say a 2N3055 NPN transistor
Connect the emitter via 47R to 0v
Feed the platform from the windsheild wiper of a 1k pot wired across the 6v provide.

The trouble is the the globe is fine for lower voice frequencies but for higher nbtv video it needs it can't be modulated this way perhaps as Steve thinks lower the bandwidth with fewer lines perhaps ...it would be fine for a slow fax system ...Both Baird and Jenkins made a matrix flat screen television with little globes i suppose that fixes the bandwidth problem spreading it out over hundred's of globes giveing them enough time to cool and heat up again .