I have a very basic understanding of how a BJT or a FET works, but I think I got the idea. They are awesome in all regards, except one: you can't make one at home without hot ovens and scary chemicals. The old triodes would be easier to make, except they need good vacuum (better than what I could make at home).

Are there any other components that one can use in the place of a transistor or triode, to control a strong current with a weak one?

I thought of these ways - they are naive ideas of someone fascinated but who barely entered the "Novice" stage of electronics:

  • make a weaving of insulated copper wire (like a tight grid) and pass other wires through the eyes of the weaving, so a voltage in the weaving (the controller) could scare off the electrons going through the passing wires (the controlled) using its electric field, like how the grid works in a triode, or how the gate works in a FET

  • put a plate (controller) between the plates of a capacitor (controlled), so when you put a current through the controller-plate, the capacitor's capacity is modified a bit so its reactance changes (I haven't given much thought to this idea, nor to how could it be used, it will definitely sound stupid)

  • put a cylindrical capacitor (the controller) around a wire (the controlled), so the negative, inner cylinder will be close around the wire; hopefully, charging the capacitor will ... oh wait... the net electric field outside a capacitor is supposed to be 0...

I already assume these ideas are wrong in innumerable ways. Could you please point out some of the mistakes? I hope I'll learn a lot this way.

Transistors and triodes are the best, most convenient, quickest and most efficient for this task. I know it very well, I don't deny it in any way. This question is asked in the spirit of "What I cannot create, I do not understand." and "Know how to solve every problem that has been solved.", as Feynman so brilliantly put it.

I'm primarly interested in control as in amplification with a gain. Switching on/off isn't really enough - I think I'd like to use this thing in an oscillator or RF circuit.

  • \$\begingroup\$ Just out of curiousity, what are you doing that you need to make your own, rather than just buying one for no more than a couple of bucks? \$\endgroup\$
    – user758556
    Commented Jun 9, 2012 at 18:40
  • \$\begingroup\$ Right. Should have mentioned. I don't need anything special. It's just to be able to make my own. Once I succeed in making something at home with crude technology, I'll happily use all the transistors available, with all the conveniences they provide. But before that, I'll be tormented by the thought that "yes, I can easily use a readily made one, but can I MAKE one myself?". \$\endgroup\$
    – Camil Bancioiu
    Commented Jun 9, 2012 at 18:54
  • 1
    \$\begingroup\$ Many of ideas you suggesting were lost with Tesla and are now patented by IBM \$\endgroup\$
    – user924
    Commented Jun 10, 2012 at 3:40
  • 1
    \$\begingroup\$ Your first idea reminded me about things like the "mono electrons", "giant magnetic resistivity", "hall sensors" and "weigandt sensors". Some of those were achieved by IBM and certainly Tesla was thinking or trying this kind of things \$\endgroup\$
    – user924
    Commented Jun 10, 2012 at 10:27

4 Answers 4


There is the magnetic amplifier. I think that it looks pretty easy to build, and works rather well.

Here's a good book about mag amps: Magnetic Amplifiers, Paul Mali, 1960, 101 pages, found in Pete Millet's truly awesome book collection. From this book:

Simple Saturable Reactor. Source: Magnetic Amplifiers, Paul Mali, 1960

"Basically, the principle of the simple saturable reactor can be stated in two parts: As magnetic core saturates, current to load increases; as magnetic core desaturates, current to load decreases." (p.28)

Another picture illustrates this:

Magnetic Amplifier Input vs. Output Waveforms. Source: Magnetic Amplifiers, Paul Mali, 1960

When you vary the DC current (control circuit), you drive the core into or out of saturation, controlling the slope if the output-to-input transfer characteristic (AC load circuit).

Note that because the underlying principle is an inductor, you can control AC load currents only. For DC applications, it is possible to control an AC current and rectify the output, using a diode rectifier and a smoothing capacitor.

Besides the rather exotic applications in early military technology, avionics or high voltage transmission, there are at least two examples where saturable inductors were used by the millions in consumer devices:

  • Post-regulation in switch-mode power supplies - actually an applicaton where an AC current is controlled to create, after being rectified, a stabilized DC voltage.

  • Circuits for correcting the otherwise distorted geometry ("cushion shape") of a picture on a CRT display. Here, the AC current to the picture tube's deflection coils is modified by a controlled saturable inductor.

  • 1
    \$\begingroup\$ Nice memory. Magnetic amplifier shoudl be powered with at least ferroresonance voltage stabiliser \$\endgroup\$
    – user924
    Commented Jun 10, 2012 at 3:32
  • \$\begingroup\$ Also known as the "saturable core reactor" \$\endgroup\$
    – Nick T
    Commented Jun 10, 2012 at 4:22
  • \$\begingroup\$ From what I've read, a magnetic amplifier must be powered by AC only. Does this mean that I can't use it to make a battery-powered amplifier for a microphone? Or I actually can? \$\endgroup\$
    – CamilB
    Commented Jun 10, 2012 at 7:48
  • 1
    \$\begingroup\$ @Camil. You will need a stable working point in DC region. So I thought that correct offset source without contemporary electronics should involve stabiliser. The ferroresonant voltage stabiliser is the kind of AC stabiliser for the task. To turn AC into DC you will need a synchronous rectifier (no electronics again) and low pass filter. But for start the ofset can be made with just a permanent magnet. \$\endgroup\$
    – user924
    Commented Jun 10, 2012 at 10:21
  • 1
    \$\begingroup\$ @Camil: "I thought I couldn't use [magnetic amplifiers] with DC. Please tell me I'm wrong!" - I have edited more information into the original one-line answer, and you will find that you can control DC - the limitation is that you need to run AC through the mag amp and you have to rectify this current. \$\endgroup\$
    – zebonaut
    Commented Jun 10, 2012 at 21:22

Acheiving real power gain is going to be tricky. Things you can do on your own without specialized equipment aren't going to yield high gain or work at the frequencies you want. Put another way, you're not going to get much gain*bandwidth product. If it were that easy, vacuum tubes, then followed by transistors wouldn't have been such breakthrus.

However, a few things come to mind:

  1. A relay, which is something you can make yourself. It only has two states, but quite high gain. You can get analog output via dithering, or PWM. One way is to set up the relay to self-oscillate. Set up the contact so that it is normally closed, then put it in series with the coil. This is exactly how old buzzers and electric bells worked. Now you've got a bunch of pulses at probably a few 10s of Hz. At this point it takes little control input to perturb the system to a different duty cycle. The control could be a separate electromagnet pulling on the contact arm or even a second winding on the same one. It should be possible to set this up to require less control power than the power that can be derived from the average switched output, thereby providing gain.

  2. A controlled light and photoresistor. Since the voltage and current thru the photoresistor could be higher than what is driving the light, this could possibly provide power gain. Maybe it can only be set up with off the shelf parts to have power gain at AC. That's not cheating, since a lot of vacuum tubes have the same issue due to the power required to heat the cathode. Perhaps at the right operating points with a low resistance LDR you can achieve AC power gain. Voltage gain by itself should be easy, even at DC.

    Optocouplers exists that definitely have power gain, but these all use phototransistors, which would be cheating according to your requirements.

    The hardest part will be to make your own light that has any reasonable effeciency. Getting the materials for a CdS light-dependent resistor won't be easy either, but I think they should be possible to construct on your own although I've never tried that.

  3. Like above, but use a fixed light source and modulate its beam by deflecting it via a magnetically or electrostatically controlled mirror. This might be the easiest way to make your microphone amplifier. The microphone could be something as simple as a sheet of aluminum foil that reflects sunlight or some other strong light source. With some lenses, it should be possible to turn this into a modulated light beam. That driving a photoresistor and a fixed DC voltage source might be enough to drive small headphones to audible levels.

  4. Magnetically controlled throttle. A gasoline engine is capable of large power output, certainly much much more than what would be required to actuate its throttle electrically via a solenoid. If you want electric in and electric out, have the engine drive a generator. This can definitely have significant power gain, although the bandwidth will be low, like a Hz or two at best.

  5. Field-controlled generator. Make a electric generator with field coils instead of permanent magnets. This has the advantage of being able to handle large amounts of input power in the form of a turning shaft. With everything just right, it is possible to modulate the large input power enough to end up with a net power gain from the control signal to the output. It may be easier to get AC gain by perturbing a DC operating point. Some early RF transmitters worked on this principle.

Basically look for anything in nature that handles reasonable amounts of power, then think how that power flow could be modulated with a lower power control signal. There are many such things if you look carefully.

  • \$\begingroup\$ Well, this dampens my enthusiasm a little. Many thanks for the explanations. I inadvertently ignored the bandwidth, I think I just assumed it's going to be fine in every case. Two more questions, please: (1) Could I use a magnetic amp, powered by the AC-like signal generated by the relay you described at 1.? \$\endgroup\$
    – CamilB
    Commented Jun 10, 2012 at 14:45
  • \$\begingroup\$ And (2): If all else fails - probably will - and I end up trying to build a vaccum triode, would it be very sensitive to the quality of the vacuum I can make? Say I bring it under 100 Pa of pressure. Would it be working? Or better: what formulae should I be looking at to calculate the triode's impedance when I know the pressure of the vacuum? I assume I need the ε (permittivity) somehow - probably measure it with an improvised capacitor placed in a similar vacuum. Am I making sense? \$\endgroup\$
    – CamilB
    Commented Jun 10, 2012 at 14:48
  • \$\begingroup\$ @Camil: I don't know what kind of vacuum is needed for a reasonably functional triode. I saw a video on youtube a few years ago by some guy in France that made his own vacuum tubes and then built a complete radio receiver. Very impressive, but he also had a lot of specialized tools, particularly a contact welder. \$\endgroup\$ Commented Jun 10, 2012 at 14:54
  • \$\begingroup\$ @OlinLathrop: Yes, that video is still up and it still inspires people. I'll eventually look into the formulae that determine the impedance of a triode as a function of the vacuum within. Probably next weekend... \$\endgroup\$
    – CamilB
    Commented Jun 10, 2012 at 18:34

Have you considered a flame diode or triode? These are vacuum tube equivalents that operate inside the flame of an alcohol lamp instead of in a vacuum. This guy has built both, and the site has pictures of his devices and descriptions of the materials and techniqued he used.

You will need a source of high voltage DC (he uses 200VDC) and you will have to work with very low currents - you will need an existing amplifier (built with regular transistors or ICs) to hear the oscillator and/or an oscilloscope with a very high input impedance to view/measure the output.


There is also this guy who has made field effect transistors out of zinc oxide. There are two pdfs on the linked page. One describes creating a simple transistor, the other describes using a further development of those transistors to make an oscillator.

The nastiest thing he needed was acetic acid (98% pure.)


I suppose you could try to rig up something with a inductive relay, a spring and a rheostat...but triodes and the many kinds of transistors are how you do it.

  • \$\begingroup\$ My bad, I should have been more clear. I'm interested in a transistor's/triode's amplification features, not just simple switching. Relays are just switches... \$\endgroup\$
    – Camil Bancioiu
    Commented Jun 9, 2012 at 21:16
  • 1
    \$\begingroup\$ @CamilBancioiu: Yes, that's why you need to connect it--mechanically!--it to a rheostat which controls the flow of the higher current. And the purpose of the spring is to allow you to have a smooth rather than digital transition as you turn up the current to the relay. \$\endgroup\$ Commented Jun 9, 2012 at 22:44
  • \$\begingroup\$ Ropes and sticks. And selsins. Dont forget selsins. \$\endgroup\$
    – user924
    Commented Jun 10, 2012 at 3:33
  • \$\begingroup\$ @dmckee: okay, this means I didn't understand your suggestion initially :) it's still unclear to me, but now I see you have a point, and I'll look into it. Cheers for the patience, and for not rejecting the weird restrictions I added. \$\endgroup\$
    – CamilB
    Commented Jun 10, 2012 at 7:40
  • \$\begingroup\$ @RocketSurgeon: crazy, so it must work. You're talking about these? The more reliance on ropes and sticks, the more I like it. Now if I could add some rocks to it... \$\endgroup\$
    – CamilB
    Commented Jun 10, 2012 at 7:46

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.