#381

[@DanielLucas-ge6rd]

Btw. I am the owner number 3 of a TEK 465B. After a litlle refurbish work it reveals his beauty. Cheers from Argentina.

[@thosewhowish2b693]

The cheeky "I'll see you later" ("oscillator") is just gold. xD

[@TheElectronicDilettante]

I have scope envy!! I was happy with my Tek TDS544 until I saw your MSO. That thing is no joke, Super cool! Thanks for the video. The Quantity and Quality of this genre of content has been lacking lately and doesn’t seem like it will improve anytime soon. So, whenever you post a new video, it’s kind of a big deal. Thanks again!!

[@wb7ond]

Check out Rigol's new 12bit DHO 900 series scopes.... Waaay more affordable, and with lots of the same type of "bells and whistles".... Of course, not in the same league as Tek...

[@aidanstanford6742]

Each time I notice one of your Circuit Fun videos in my suggestions, I watch it. Very informative and entertaining, definitely one of my favorite mini-series on KZbin!

[@TheStuffMade]

Impressive FFT on your scope! Some time ago I had a look at an old LT application note on a very low distortion oscillator circuit using various high end LT parts. I believe it achieved something like 0.03% (-70dB) distortion, but then again these days you can get a $9 Apple USB-C headphone adapter that will do 0.001% (-100dB) distortion @1kHz, 1Vrms output with an output impedance of just 300 milliohm.

[@JaenEngineering]

​@@jim9930I have the same scope and as far as I'm aware it's 8 bit only on the 4 analogue channels and 1 bit on the 16 digital channels (if enabled).

[@TheStuffMade]

@@jim9930 It's an interesting method to measure the fundamental and then at the same time notch it out and measure the harmonics with increased resolution, but I guess it's not much use if you're doing a THD+N/SINAD measurement. I've found the Digilent Analog Discovery to be a decent budget solution with it's 14 bit scope having differential inputs and 14 bit AWG doing 100/125Msps and effectively good up to 15-20Mhz bandwidth. I've designed both an Impedance Analyzer and an Audio Analyzer using that device.

[@W1RMD]

Congratulations on 200k subscribers, first of all. It’s VERY well deserved. Also, thanks for doing a follow up on this video. I love that that scope screen splits in half so that the spectrum analyzer and scope can be viewed at the same time. Take care and toota-loots (yes, I'm an electronics for the inquisitive inventor fan too), oscil-later, 73, and all of that stuff.

[@SatyajitRoy2048]

That MSO is a beast no doubt. Your video are very definitive and clear on the topics. Wish more frequent videos come if you have enough free time to play around.

[@ptronix]

Hi Alan, watched both videos with interest. Last year I went on a quest to build a low distortion sine wave generator. Started off with the usual Wein bridge which was OK but the distortion was way higher than the source in my hp 8903 audio analyzer. So then moved on to what was called a 'State variable sine generator' uses 4 opamps, and a bulb for stabilisation. The bulb seems ultra critical. Needs to be fairly high voltage with low current for best results. I started off using a TL 074 and latter moved on to the more exotic OPA 134. Both chips were very good. The final circuit measures around 0.0013% distortion, measured on the 8903 and on some computer software which also displays the spectrum. The harmonics and the biggest (2nd) was way down at -95 db with respect to the fundamental. In fact the distortion is so low my Keithley 2015 has problems measuring it! It is now my default signal source when measuring amplifiers. I did several videos on this project on my channel should you be interested. Best wishes nick

[@davidt7317]

Nice clarification, Allen. Well done as usual.

[@KK-rg3nj]

man i love when ytbers actually listen to comments especially in cases like these. Thanks for the video, the odd harmonic and symmetric distortion was something i knew of but this was the first time seeing it actually haha

[@PapasDino]

It's interesting that the scope shows a bit of distortion at the top/bottom even with such a small amount of distortion. Seems like other distortion would come into play before flat-topping came into the picture...or maybe flat topping is what comes first? TNX Alan for the follow on demo. 73 - Dino KLØS

[@JoseSilveira-newhandleforYT]

Very informative follow up, Alan! If this was an audio amplifier, just the visual quality of the sine wave in the scope would allow adjustment for probably less than 0.1% THD (assuming a good scope and a good amplifier design, of course), which is very good for audio quality.

[@RadioWhisperer]

Another great video Allen, thank you for doing these. I review your series on various topics when i need a refresher. Your content is a cut above the rest. I am stunned by that scope, only word that comes to mind is wow!

[@johnyoungquist6540]

Touchy to adjust, wouldn't take much to get it to fail. The Wein bridge with a little light bulb (T1 size or smaller) works great and is stable over a wide range of conditions. The bulb shows remarkable AGC behavior. I thought it was dumb until I tried it decades ago. The current is very low it doesn't light . Low distortion unlike zener designs. I made a single transistor version of a tuning fork oscillator with the bulb in the emitter. After all it's how HP got its start. Make a new video and compare. Great video as they all are. JY

[@andymouse]

Exactly what I have been playing with ! the bulb is awesome, still thermionics can help out ! best I've seen using sound card and ARTA is 0.005% THD with my version of Dr Hewlett's work and my own PCB.

[@jamesmorton7881]

I really appreciate seeing elegance combined with simplicity. Wish i had found this circuit, Uh like twenty years back.. A related slick passive is to use a high Q RLC to obtain a healthy VPP above the rails for modulating a fiber optic laser carrier. DDS DRIVING A CLASS B into the peaking before †he laser modulator that requires around 50 VPP. Also a multiplying DAC for precise amplitude control. All great for sycncro resolver excitation applications. Fiber optics are used to make all dry ( no wires ) hydrophone arrays. Amazing to see the elimination of many hardware stages in software. I worked on a VSP at Litton Guidance. vertical seismic profiling using a fiber optic array down hole. for oil deposit mapping. Interesting is using shift register and XOR TO make random noise.

[@isettech]

Back in the day, an actual incandescent light bulb was used in bench equipment to do the job of reducing harmonic content by preventing clipping. When the amplitude increased, the incandescent filament would warm and increase in resistance. This worked very well with the tube equipment. For the Op Amp version, a Light Dependent Resistor, coupled to an LED can perform the same function, but with the change of the LDR reduces resistance instead of increasing with increasing voltage. The AGC function can keep the oscillator out of the clipping region by riding the gain a few DB below clipping for good signal to noise and very low THD.

[@AstroSam66]

Very nice video, thank you! I would use a multiturn pot, that would make it easyer to find the sweet spot.

[@JohnUsp]

Excellent explanation and demonstration. Thanks.

[@DanielLucas-ge6rd]

Very good video and well explainded. Thanks for shearing!!

[@NeedleBender785]

Great video! Thanks for all you do...!!!

[@xDR1TeK]

Alan!!! Make more videos!!! Plz. Build an amplifier next to this, a buffer, then filter the harmonics for us. Wanna seeeeee. I do a lot of circuits, seeing you do it is more gratifying.

[@aduedc]

Great clip again, Lets double the caps ( by putting the same cap in parallel with each cap) and divide the resistors by 2 (by putting the same resistor in parallel with each resistor) see if we can acheive higher or lower THD. This should not change the frequency because the product RC value is the same. Also, I wonder how Caps and resistors mismatch effect the THD.

[@mnmmnm6718]

U are a true GENIUS...

[@VoidHalo]

I'd love to see a video about isolating harmonics from an astable multivibrator the way the original circuits were originally intended to be used. Maybe even touching on the history a little bit. I could never find a description of EXACTLY how they isolated specific harmonics from the squarewave output of the multivibrator. I always assumed they would use a bandpass filter and then amplify the output. I tried it myself, but the passband on my filter was always so wide that other harmonics got into the output.

[@SuburbanDon]

Thanks for adding this.

[@Gameboygenius]

Ok, this was interesting as I suspected and raised more questions. Firstly, what is it that we're actually trimming for? I suspect that we're actually trying to match the depth of the notch to the (open loop) gain of the op amp. Make the notch a little too deep and oscillation stops. A little too small and the amplitude goes too high. BUT the implication of this (I think?) is that the two T's are actually slightly detuned from each other, compared to the ideal R/C equivalence, to bring down the gain in the notch slightly. Am I understanding this correctly or am I talking nonsense? Also I noted, especially at 4:34 that the 6th harmonic is almost completely nulled out. So it's not just a split between odd and even harmonics. Lastly those comments with someone trying to flatter you with a profile picture with their butts out are spam bots. They either copy a comment from a previous video or use AI to generate a comment. I'd suggest deleting those.

[@alaricsnellpym]

I'd worry that, as this design requires you to get as close to oscillation dying as you can to optimise THD, that if I tuned it too close to the limit then one day it wouldn't oscillate because the ambient temperature was wrong or something. As the distortion seems to be clipping as the op-amp aims to maximise its output dynamic range, I wonder if some kind of AGC circuit could be used to constrain the output swing, to keep some safety margin away from the op-amp limits. Hmm...

[@w2aew]

Yes, you’re defining to just meet Barkhausen criteria, so not really practical or robust, just something fun to play with.

[@w2aew]

Exactly right.

[@alaricsnellpym]

@@w2aew I live for "fun to play with" 😁

[@TechneMoira]

Thanks for this follow-up. I was also wondering what the effect would be on the total number of harmonics if one would work with half the capacitance on the highpass filter... or conversely, on the lowpass filter. Could you find your way to demonstrating that when you have some time please? Thanks a lot for the very clear explanations P.S. by half the capacitance I mean as half the capacitance on the other filter or vice versa

[@bobdoritique7347]

I try your schema, very nice.

[@WolfgangMahringer]

It would be interesting to see how supply voltage influences THD here...

[@w2aew]

It will likely have an effect.

[@rovhalgrencparselstedt8343]

I imagine that with a multiturn pot you could get the distortion down alot more before/without killing the oscillation completely.

[@mikaelstrom1114]

Interesting circuit. Is the temperature affecting the stability of the THD-adjustment? I mean, is there a chance that oscillation spontaneously stops if adjusted to close to the limit?

[@w2aew]

Yes, quite likely.

[@shean4]

Thanks for the video. Are you able to extract the harmonics from the scope in a data table and then use those data points to calculate the THD and compare that with the multimeter value? :)

[@mikesradiorepair]

Yes you could. A video I did several years ago shows the formula and how to calculate THD. kzbin.info/www/bejne/rWa4fK2aprd0p6cfeature=shared&t=2784minutes The part of interest on THD starts at 46:25 minutes.

[@w2aew]

It would be possible. You’d also have to match the number of harmonics included for each measurement.

[@shean4]

@@w2aew Yes, I was wonderng what bandwidth the Keithly used for its measurement of THD.

[@w2aew]

@@shean4 I believe the 2015 can measure harmonic components to 20kHz.

[@waynegram8907]

Why doesn't the FFT or Spectrum analyzer read THD? it should has a THD readout in the menu settings

[@henrikstenlund5385]

Thanks for sharing this. The common problem with oscillators is that they are sometimes hard to get running and if one bulds an amplifier, it tends to oscillate if given any chance. Have you studied this circuit regarding the reliability of starting? Is it critical to temperture / part tolerances?

[@w2aew]

I have not studied it in those terms - just this introductory video....

[@technishn]

Great videos. Does a non rail to rail op-amp like a 741 have the same order of distortion?

[@w2aew]

All op amps will distort when they reach their output voltage range limit.

[@joopterwijn]

I know the THD is low, but in my mind I was placing the same filter behind the opamp output. Wondering 💭 if you even could eliminate it at all. I know overkill just playing in my mind.

[@Swamp-Fox]

Oscillator indeed!

[@NeedleBender785]

Good morning!!, Had a technical question. This twin tee looks like a band pass filter on the vna, in theory would you be able to use a crystal placed instead of that and get a sinewave without changing the rest of the circuit?

[@w2aew]

You certainly can make a crystal oscillator - the circuit will vary a bit, but they work. However, crystals are typically not available at these low frequencies.

[@viperwizard491]

Jim Williams achieved 3ppm with Wien oscillator. ~500x improvement

[@w2aew]

Not surprised at all Jim Williams was a true analog wizard.

[@Amir-kz6yq]

Interesting demonstration. Has anyone else tried to build this oscillator? I both simulated it in Ltspice and built it on a breadboard without any luck. It just outputs a dc voltage at vdd/2.

[@chasson321]

Could you do a video on a circuit that takes high frequency dc pulses above and below to a 60 hz sine wave?

[@w2aew]

Not sure what you mean by that.

[@chasson321]

@@w2aew I’m looking to replicate a Don Smith magnetic resonance system. The device uses a double Tesla coil and generates a double DC spike like a sign wave with a high voltage and high frequency around 200,000 cycles per second. I understand how to get the voltage down but unsure how to use a resistor and capacitor to smooth out the spike and lower the frequency to 60 hertz. Any thoughts?

[@brianclements1014]

I can’t get my breadboard twin tee to oscillate! Built it several times using o0.01uf caps, and 20k ohm resistors. R/2 with two R in parallel, same idea 2c. Have used TL081 and ½ LM328 op amps. Voltage from 5 through 12 volts. No oscillation. Can’t understand why

[@w2aew]

You may need to make the R/2 leg adjustable...

[@brianclements1014]

@@w2aew I’ve sort of tried that - how critical is the R/2? I’ve use Rs from the same resistor strip, and they are quite close to the same value, hence my R/2 should be reasonably accurate. But I’ve tried a ten turn pot as well that passes through my desired 10k point, but no joy. Maybe I’m too coarse in adjustment. I’m using the circuit from IMSAI guy , and note he has a 110k R from output to the + input, and a 1k R from + input to ground. How critical is that I’m thinking. I note that you don’t have that in your circuit, but instead lift the voltage a little at the + input to op amp. …… this should be simple, I’m usually working with RF! Maybe there’s an issue with my op amp choice (read availability). I’m in a remote location. Just want a nice 800Hz - ish sine wave oscillator to use for my CW practice. VK4BAC, but unlicensed for XW region.

[@ghlscitel6714]

What is the total harmonic distortion of a soundcard output measured with the same setup?

[@w2aew]

Probably quite good

[@ghlscitel6714]

@@w2aew Would be interesting to see in comparison.

[@andymouse]

I have a cheap one and measured in 'loop back' better than 0.002% Edit :- using ARTA software.

[@AlexejSvirid]

Tektronix pays you for using it's equipment? I'm envy. ;-)

[@w2aew]

Using, training, presenting, demonstrating, problem solving, etc. - best job I've ever had, for sure! Many of my "work" videos are on the tek.com website and Tektronix youtube channel. This one here, it's for me....

[@W1RMD]

A man who loves his job never works a day in his life.@@w2aew

[@TKomoski]

My guess this person doesn't know you @@w2aew

[@AlexejSvirid]

​​@@w2aew Here is the God's word written by king Solomon, Ecclesiastes 3: 22: "Wherefore I perceive that there is nothing better, than that a man should rejoice in his own works; for that is his portion: for who shall bring him to see what shall be after him?" Enjoy your work, Alan. :-) Unfortunately, not much people can do the same. The problem is Devil runs the world. He is lier and murderer. That's why we've got the Gospel about the God's kingdom. Jehovah would put everything in order. The dead will be resurected and we'll meet our beloved ones again! :-) Edit: hope your employers don't try to envolve you in a crimes.

[@AlexejSvirid]

​@@w2aewEdit: hope your employers don't try to envolve you in a crimes.

[@wiwingmargahayu6831]

zamzam water

[@stevenbliss989]

How about some POSITIVE FEEDBACK to keep it oscillating with even lower distortion?

[@w2aew]

Not sure how positive feedback will lower the distortion.

[@stevenbliss989]

Not directly, but expand the adjustment window in which it will keep oscillating, which MIGHT find a spot better than 0.16% THD.@@w2aew