RU-vid decided I needed to see this video. Even though it seems to be six years old, it showed up on my feed. It brought back memories of when I was repairing audio gear in college. Our HP generator got a little unstable and I decided to crack it open. Imagine my surprise when I ran across a large (for something found inside a piece of gear), odd looking light bulb. Being young and dumb, I assumed the bulb was bad and ordered another one. I forget what was bad, but it's not a spoiler to report that it wasn't the bulb. Such a simple, elegant means of gain control. Great memories.
I too as a kid wondered why there were light bulbs in oscillators! One neat trick that I used ( I got it from an old book) is a radio squelch circuit made from two ten ohm resistors and two number 49 light bulbs. It's a bridge. I used it for years on a 6 meter SSB receiver. It goes between the speaker and the receivers audio out connection. You increase the audio gain until the lamps heat up enough to balance the bridge and then the speaker goes silent. Any signals that come in unbalance the bridge and un-mute the speaker.
One of my earliest electronics projects was a Wien bridge oscillator using a 741 op-amp. That was some forty or so years ago but I somehow clearly remember the struggle to get it to work and how I finally eventually succeeded. I had decided the problem was that the 741 could not drive enough current to warm up the lamp enough and I couldn't find a smaller lamp - the best I had was a bulb form a Christmas light string. In an attempt to solve that I put in an emitter follower stage to boost its current output capability. That could then easily drive the lamp but no matter how I adjusted it I could tell just by hearing that it was severely distorted. I then solved it in a different way. I replaced the lamp with an LDR. This acted as the gain controller. I then used the emitter follower transistor to drive the lamp which illuminated the LDR acting as the actuator for the gain control. With a lot of tweaking I finally got it to work. Many years later when I had access to an audio spectrum analyser at school I tested it and was pleased to find out it actually had very low distortion.
I never had children of my own, my wife has two grown kids from her earlier marriage, but had I had my own daughter I would have liked her to be very much like you.
Feedback loop oscillators reach stabilization when their closed loop gains reach 1. This is often done when the active device goes into saturation and the gain at the primary oscillation frequency is reduced while harmonics are increased. The use of an incandescent lamp, a device with resistance that has highly positive temperature co-efficience, allows loop gains to be reduced without the active device going into saturation; meaning oscillation reaches stabilization with much less harmonics (all active circuits are non-linear to some degree, even with heavy negative feedback). That is the primary benefit of using a light bulb to control oscillation. The other benefit is to assure reliable oscillation start-up because the resistance of the bulb is very low before voltage is applied. The ability to tolerate power supply voltage variation is a secondary benefit. By the way, Meacham described the use of lamp in a bridge to control gain in 1937, two years before Hewlett used it in his first product.
Thank you for the reminder. I have used one of those tiny indicator lamps in a reference 1khz oscillator that provided the synchronising bleep from the sound mixer to the audio recorder at the beginning of every take shot on a 16mm film camera. Everything was powered from the camera battery. I was amazed at the stability of the signal level that this simple technique provided.
Between you and Dave from EEVblog I have rekindled my interest in building stuff! Thank you. I was the only person at school who was interested in electronics, as a result there was no point in teaching this. I think I will be making my shack more focused toward construction when I move later next year.
Hi Jeri, Just seen your video regarding the Wien Bridge oscillator which reminded me of an interesting story relating to what I believe was the first use of a light bulb in this circuit. My story starts in 1967 when my college class mates wanted to build an audio generator as part of the practical element towards out final college exam. Out tutor, the wonderful Mr Dave King convinced the class to use the Wien Bridge circuit which used a glass thermistor in the feedback (pre op amps) circuit. Unfortunately, I was the only one in the class who had a glass thermistor and were too expensive even to purchase in bulk. Therefore, not wanting to abandon the project, Mr King went home that night and designed a circuit using a light bulb. As his design was so good, he published the project in one of the UK's leading electronic magazine, Practical Electronic. I still use my original audio oscillator in my workshop. It used to have a label on the front “May 1967). Therefore I believe our tutor was the first UK engineer to design this type of circuit. Regards Geoff.
I do not build any circuits like this, but it is great to watch your videos to help me stay up on the theory behind some of these circuits. Keep it up!
nice to discover your channel Ellie. I remember when I was a kid, I would be doing descartes 5th degree equations with either hand while watching TV, and I would also be thinking about op-amps with lightbulbs. :) coincindence.
@cubemike99 The lamp is a 3V 15mA 'grain of rice' as used on model trains, get it on EBay. Resistor at the top of the diagram is a 1k pot. R1+2 is a10k dual pot with a 560R in series with each section. C1+2 100nF for about 200Hz to 2Khz. Opamp is a 741 or TL081.
Nice video, I just had the experience of building one for an am transmitter circuit! Took an annoying amount of tweaking to get the thing to work right
Nice to see the good old Wien bridge oscillator around. Like all oscillators of this type the main task is to overcome the losses to achieve a non decreasing amplitude of the sine wave. In the transfer function of the closed loop you easily spot out that the linear term in the denominator must be zero to reach a stable amplitude of the output. Here´s where the light bulb comes in. The bulb has a resistance which is dependent of the current flowing through it. The more current the more resistance. Filament heats up slowly resistance increases slowly. The "slow" behaviour is important to avoid a resistance changing according(along) to the sine wave. But you have to watch for the proper DC bias point because the bulb does not behave the same way in the whole current range. You have to spot out the highest slope of the resistance vs. current function R(I). Not every bulb works, if the slope is too flat you can´t compensate the lack of amplitude with the gain of the amp, because the gain would not grow enough due little resistance change of the bulb. Once you have figured out how to set the DC operation point of the bulb you also know the current which flows through the resistors for the negative feedback loop. That's one equation. The gain which is set by the negative feedback resistors is also a parameter for the oscillation criteria. You need just the right amount of gain to get the circuit oscillate with a stable amplitude. That's the other equation. Now the linear equation system can be solved for the two negative feedback resistors involving also the resistance of the bulb. For the frequency determining part, every circuit with the transfer function G=k/(as^2+bs+c) can be used it must not be necessarily an RC combination but also LCR or a crystal or whatever works. As long you compensate for the linear term "bs" this thing will oscillate sinusoidal.
Thank you!!!!!! So long I wondered why that light bulb. I searched halve the Internet but everyone besides me seemd to understand. You finally explained it so clear and easy. I subscribed and hope to learn a lot more.
There's a detailed analysis of the Hewlett Packard light bulb Wien circuit by Jim Williams, in the book www.amazon.ca/Art-Science-Analog-Circuit-Design/dp/0750670622
I love the elegance of this circuit! A light bulb as a thermistor. I remember seeing this one in some op amp circuit books and I too wondered what the hell a light bulb was doing in there. Now I know. :) Great video Jeri... thanks! Keep On Hackin!
Thorens used a Wein Bridge Oscillator in one of their old belt-drive turn tables, to drive an AC synchronous motor and still be able to have a pitch control and easy electronic speed adjust without having to move the belt to a different motor pulley. This was in the TD 124 or TD 125, I forget which. An excellent turntable if you put a good tone arm on it by the way. Many of the other Thorens turntables weren't all that great (and had flimsy integral tonearms) and the 1980s models with DC motors tended to have speed drift as the motor warmed up or aged; synchronous motors are really much more stable over time.
Great demonstration! Actually the first resistance tuned oscillator with a non linear element (incandescent lamp) was Bill Hewlett's brainchild alone- It was his Stanford University thesis project in 1938. Dave Packard was on the east coast and had just been married. The partnership happened shortly thereafter, although they had met and been good friends throughout their time at Stanford. The other BIG advantage to using a lamp was amplitude stability over a very wide frequency range!
Thanks Jeri. I'm going to build a Wein Bridge Oscillator and test it through a VCF and VCA. What I liked about your video were the harmonics which aren't usually wanted but are desired for subtractive synthesis. I'll l try it with a thermister and a photocell. BTW, nice scope!
Light bulbs are still used in a lot of circuits. I took apart a giant Compaq robotic DLT tape library and the motor control circuits used a whole bunch of small (~50w) quartz halogen lamps in the power supply. Guessing part of the bus voltage shunt to bleed off voltage when the motors are decelerating.
I'm sorry. I've just come back to this video. Revisited, if you like. I GET IT. Thank you so much. Of all the videos out there trying to explain Wien bridge oscillation, *THIS IS THE ONE FOR ME* This is *MY LIGHTBULB MOMENT* Thank you so much lady. I am so excited right now, I need to go lie down. LOVE IT! FREAKIN' LOVE IT.
It's not just the non-linearity of the light bulb - it's the fact that it's time constant is long enough that it varies hardly at all over the period of the oscillator.
Excellent documentary, Jeri! I never knew why they'd use incandescent bulbs either, until now :) It reminds me of the use of 250V neon lamps as surge arrestors..
I never tought the light bulb to be lame in any circuit. Many times it was preffered instead LEDs. Since 8, when I build my first "lab" AF oscillator that was a Wien bridge with transistors only, I many times, used the Wien bridge as a stable and clean oscillator, with opams or transistors or even tubes even in RF (LC configuration with L or C variable). And, at first I wondered too what that bulb is doing there but, after opening the circuit, redrawing it and reading the explanation I understood their role in keeping the circuit stable. Strange is that Internet now (re)discover this circuit. Maybe soon you will find that existed many other configurations of oscillators RC, LC or even RL that were used before digital ones that are today.
Great idea, that light bulb. Thanks for bring it to our attention. But I guess you can have a pick detector on your oscillator and then adjust the resistor that sets the gain of the oscillator. If C1=C2 and R1=R2 than your feed back factor beta is 1/3, and you need a gain of 3 out of your opamp, which means the ratio of R to Light bulb resistance should be 2. So, using a N type MOSFET biased in the ohmic region and a constant resistor can be replace your light bulb. Now you can make a peak detector and a feedback network to adjust the gate voltage of the NMOS, in order to get low phase noise. This can also cancel out the pushing (frequency change as result of supply voltage change.)
That or any other diode-based 'clipper' will work and I've done it too but the THD specification will be worse than the light bulb, provided that the light bulb has a time constant that is a lot longer than the output frequency.
Great insight on a tiny non-linear resistive element :) I wonder about choosing the right parameters (nominal voltage and cold resistance; resistance-temperature characteristics) when designing the oscillator. Maybe Mr Carlson has something to say about it?
I'm in love with your brain :) Personally I find it very useful to know the physics of basic materials used in electronics( light bulb filament heats up increasing resistance, etc. etc. ), not only does it help in understanding of how circuits work, it helps you find new implementations as well. P.S. You're such a nerd
My General Radiotelephone tube-type CB radio has a feature called "Tone signal". It sends out a warbling tone to alert others via the transmitter. It uses a wien bridge oscillator with a neon bulb as part of the circuit - no transistors or IC's!
I'd kill to get my daughter as interested in electronics as you are. I try all the time to get her into it, and she can hardly wait to change the subject. : ( Keep it up,Jeri!
A guy I know exploited the same property of a light bulb (increasing resistance with increasing current) to make a guitar compressor effect (just by wiring it in series between the guitar and amp). The cool thing is that it lights up when you play hard enough. If you wanted to do it yourself, I'd probably add a pot, though, since guitar pickup voltage output apparently varies by quite a lot from pickup to pickup.
Mateo Aboy has a video "Wien-Bridge Oscillator: Amplitude Stabilization" which uses two diodes for stabilization. She refers to the lightbulb method as the HP method for obvious reasons :)
It is said that the sound of the vacuum tube is legendary and the most desirable sound. A light bulb can be considered a single-element vacuum tube, so this oscillator is both a hybrid circuit and it has that Tube Sound that makes it so good.
That was wonderful! Good to see you doing these kind of videos again. The BEST part of this one was your "little helper". I gotta make one of those. The circuit was pretty cool also. Just kidding. I never knew that a light bulb in a circuit was doing what you showed. Thanks.
There is an alternative version with 2 op-amps that gets around the fact that the common mode rejection ratio of most op-amps is quite non-linear. The theory of opperation is the same. The second op-amp gets the RC circuits and drives what in your circuit is the ground end of the lightbulb. It does about a decade better on the distortion values. This makes it good enough for testing audio circuits.
0:32 I remember a similar feeling when installing 10base2 network cards back in the 90s. There was a neon lamp attached to the BNC. These were for "modern PCs" and they were using something akin to a vacuum tube. Of course, later I found out why; the hard way. Those 10base2 NICs had a chassis ground. One place had bad electrical wiring. When I plugged in the BNC cable, I got zapped with 60 VAC and scared as the neon lamp flashed. But it did it's job as a GDT and saved the PC. Not lame at all.
I had the same experience, practically, in 1994, putting 10base2 into an IBM PC AT. Office had reverse-wired neutral and hot. I got a strong jolt. The PC survived but the network card did not. (This with the PC powered off but plugged in to mains).
Good demo of dynamic amplitude control. but with the light bulb in the circuit the sine wave appears to have some harmonic distortion: + peaks appears sharper than negative peaks (rounder).
My friend has a collection of old HP products and yes, the old Wien Bridge is quite interesting. There's a write-up on the HP8903 distortion analyzer from the 80s in an old issue of the HP Journal, well worth reading to see how analog oscillators have evolved. By the way, ever seen distorting resistors caused by its voltage coefficient? You see these things when you build oscillator/analyzers down to a few PPM nonlinearities.
I used a #1869 light bulb from Ratio Snack for a WBO, and found that it was 'microphonic'. The crimps on the ends of the filament were loose, and were ok when it was glowing since the hot filament wire expanded. But when the filament was just barely warm in the WBO, the ends made poor contact. Remember that the lifetime of this light bulb is extremely long because it doesn't get very hot. Thanks for the WBO explanation. I love a female who can talk tech like the guys!
The AF has gone to these, but I still get by with the MSO4104. These mixed signal scopes are great when you do not need any more than 16 digital lines. It beats lugging around a huge HP rack mount logic analyzer.
One thing I didn't see mentioned here which I think should have been is resonance. The diameter of the bulb filament winding relative other parts is critical in some cases because of charge transition from part to part. This will help you prevent making those pesky artifacts for one, or you can apply such knowledge to making the first basic faster than light transmitter which used a heater coil (it's patented, search it). The key is physics, geometry, and ultimately the flower of life pattern which if used as a gauge for resonance among differing radius points, will also give you knowledge on predictive circuit building so it comes out right the first time. Eliminating thermal curve between caps of any kind (everything is a cap to some degree) will prevent burnouts and make cleaner waves.
I think I have the same meter that you have sitting there in the background. Love that meter. I just watched your scanning electron microscope video -- very cool...I'm very jealous. :)
If the sine wave goes out of the scope screen, it shows more harmonics, as the oscilloscope thinks that the sine wave is actually clipped? I am not sure if the same is with your scope, but I remember, I had a low cost scope and that had the issue. So, if the sine wave is out of scope screen, I had to reduce the vertical gain (voltage) to bring within the screen (and hence scope effectively sees the entire sine wave).
i am wondering if i can take a few old 5 LMB directv dishes and use them to transmit signals..or use them to pick up other signals other than Television.
If you're speaking in general terms, there are materials which work either way. Some become more conductive with temperature, and some become more resistive. Look up NTC and PTC thermistors and the difference between them for more info.
We did this as a lab in Cambridge. We were given the components, and allowed to make any measurements we chose. Then we had to predict the frequency and voltage of the output. After that we built it and measured the output. My prediction was pretty close.
Could you choose a light bulb with a higher internal inductance to help bat against microphonic characteristics or would this be where a SS equivalent circuit becomes easier/more beneficial?
I've built a few oscillators using both the XR2206 IC and the op amp with light bulb method. In my experience the XR2206 circuits aren't as reliable as the op amp/light bulb circuits! Perhaps that's why the XR2206 seems to have been discontinued.
Hi Jeri! That's really great you showed that to me! I'm not even started to build the guitar effect I got in my mind from you one bit square wave demonstration! I'm not forget it but I don't exactly remember how it gone then but I have the visual view in my mind how the wave changed then when changed the clock! That's super cool too but the light blub is monophonic! Did you said it can changed with different resistor? I'm wondered use light blub somewhere! Thanks for that! All the best! Assa!
Jeri, you could discuss the fact that the oscillator is not always stable and any change in frequency for example causes the amplitude to ripple for some time before it settles at a new value.
It is a pretty wired effect of you put unfiltered DC (right past the rectifier) in those things. I was working on a circuit and had some bad caps in the ps so they weren't filtering.
What if we use PTC (positive temperature coefficient resistor) instead of the bulb? When Wein bridge was discovered, PTCs were not popular. But these days, they are used for many different applications (current protection).