Transistors necessary?

[ltiu]

I have been reading about boost circuits and constant current circuits. I notice they all involve transistors.

Are there ways to build boost and constant current circuits without the use of transistors?

 

[lctorana]

Short answer is yes, transistors are necessary.

By "transistor", I will assume you include FETs, MOSFETs, SCRs, unijunctiions, etc, and integrated circuits which contain transistors, in that definition.

A fascinating theoretical exercise.

Boost:

The answer is no practical alternative.

All electronic* DC voltage boosters rely on the conversion to AC, step it up with a transformer or some kind of flyback inductor, and (usually) convert back to DC. The initial DC:AC conversion is achieved by some kind of 3-legged semiconductor component, but can be done with valves (vacuum tubes), gas-discharge bulbs (e.g. neons), or a mechanical chopper (like a doorbell buzzer or oscillating relay) None of these alternatives will fit in a flashlight!

* there are also non-electronic means, such as rotary converters (genemotors etc)...
​

Constant Current:

The answer is a qualified no. There is an alternative, but it's not a great one.

Let's go back to the 1920s for a moment. If you needed to perform current regulation back then, you would use a device called a barrettor, a specialised development of the incandescent light bulb, which within its design current region, gives tight current regulation.

But ALL tungsten filament light bulbs exhibit this characteristic to a lesser degree.

So you can always replace a dropper resistor with a bulb of the appropriate voltage and current design values.

But this will give poorer regulation than a transistor-based current regulator, but better than a simple resistor. And it only works if you have enough voltage headroom to play with.
​

That was fun!

 

[matrixshaman]

Most flashlight drivers contain tiny IC chips and on average I'd say a very small IC has roughly 50 to 100 transistors. As Ictorana pointed out there is no reasonable way to do this without the transistor at this time.

 

[ltiu]

The reason I posted this thread is because I wanted to play around with (build) circuits that can boost voltages or regulate current with only simple resistors, capacitors, potentiometer, trimmer, etc. Essentially large simple components.

And not having to go get an IC chip to do it.

Thanks for the great answers.

The IC transistor is really a great invention. Without it, a lot of things we take for granted won't work.

 

[Mr Happy]

The reason I posted this thread is because I wanted to play around with (build) circuits that can boost voltages or regulate current with only simple resistors, capacitors, potentiometer, trimmer, etc. Essentially large simple components.

And not having to go get an IC chip to do it.

Then you are golden. Just add a few ordinary transistors to your armory and you can play. If you want to boost voltages you will not be able to avoid coils or transformers though. Although you can theoretically boost voltages with just capacitors, you will not get enough current that way to drive an LED.

 

[lctorana]

Although it has absolutely nothing to do with flashlights, would anyone like to see pictures of a barettor in operation, a genemotor, a gas oscillator tube, etc? Happy to post if anyone is curious to see what this stuff looks like.

 

[lctorana]

Although you can theoretically boost voltages with just capacitors...

Just for my education, and totally ignoring practicalities,


How?

 

[ltiu]

Although it has absolutely nothing to do with flashlights, would anyone like to see pictures of a barettor in operation, a genemotor, a gas oscillator tube, etc? Happy to post if anyone is curious to see what this stuff looks like.

Please please please.

 

[PhotonWrangler]

Although it has absolutely nothing to do with flashlights, would anyone like to see pictures of a barettor in operation, a genemotor, a gas oscillator tube, etc? Happy to post if anyone is curious to see what this stuff looks like.

I've never heard of a barettor until just now, so I'd like to see one. I've used NE-2 neon bulbs to build relaxation oscillators though. I think those qualify as the only semi-passive oscillators. They produce a striking ramp waveform.

 

[Mr Happy]

Just for my education, and totally ignoring practicalities,


How?

I suppose I should be clear that I meant using just capacitors as the energy storage device and not using inductors of any kind. I understand that such circuits are used to generate the high voltages needed in flash memory sticks. I think they work on the voltage multiplier principle with capacitor/diode ladders. Some transistors are needed too, of course.

 

[lctorana]

I suppose I should be clear that I meant using just capacitors as the energy storage device and not using inductors of any kind. I understand that such circuits are used to generate the high voltages needed in flash memory sticks. I think they work on the voltage multiplier principle with capacitor/diode ladders. Some transistors are needed too, of course.

Still can't see how that could work.

(I thought those circuits used tichy little SMD inductors.)

You see, when you make or break an inductive circuit, you get a + or - voltage spike, the height of which depends on the resistance in circuit. We all know that.

But when you make or break a capacitive circuit, you get a + or - current spike, which I can't for the life of me work out how to use for voltage multiplication purposes. If you add resistance to turn current into voltage, the current spike disappears.

It's probably obvious, but i just can't see it.

 

[Mr Happy]

Still can't see how that could work.

(I thought those circuits used tichy little SMD inductors.)

You see, when you make or break an inductive circuit, you get a + or - voltage spike, the height of which depends on the resistance in circuit. We all know that.

But when you make or break a capacitive circuit, you get a + or - current spike, which I can't for the life of me work out how to use for voltage multiplication purposes. If you add resistance to turn current into voltage, the current spike disappears.

It's probably obvious, but i just can't see it.

Some may use inductors, I'm not sure. Engineers are always free to choose the optimum circuit design.

The basic operating principle of capacitor-only circuits is to charge up a bank of capacitors in parallel and then discharge them in series, using an arrangement of switches to achieve this. The switches can be diodes, transistors, spark gaps, discharge tubes, or various combinations.

Here's an example of a low voltage IC: http://www.maxim-ic.com/quick_view2.cfm/qv_pk/1017

And here's a more spectacular example using diodes and spark gaps: http://www.electricstuff.co.uk/marxthree.html

 

[lctorana]

Ahhh.... Now I geddit.

So if you want 20x multiplication you just throw your 20PDT switch. Which itself can be an on-chip silicon component. Or 21PDT for a self-excited oscillator.

And Mike's Electric Stuff? You've probably just linked to the one and only page on that site I hadn't read...

(Edit: Just read up on the "Pointolite" lamps on the Lamps page. Unfeasibly complicated to operate but fascinating.)