#622

Episode 622
Use harmonics to get higher frequencies or turn a square wave into a sine wave
Be a Patron:
/ imsaiguy

[@oldblokeh]

Don't forget that if you start with a sine wave you can always stuff it through a schmitt trigger to get a square wave and hence all those lovely harmonics.

[@WistrelChianti]

Cannot like this hard enough! Thanks! Really interesting. Total bonus unexpected learning there!

[@VeryMuchBlessed]

Very interesting good information. Was good to see the process refined down to the sine wave. Thanks much.

[@gudgejm]

this video enlightened me. you explained exactly what i was curious about, thank you SO MUCH!!!

[@herbertsusmann986]

Yea, odd harmonics are not supposed to be all the same. Falls off as 1/n I believe. Can easily check by looking up the Fourier Series for a square wave.

[@ChrisSmith-tc4df]

That frequency multiplication technique is often used in Rb and Cs references to create microwave frequencies by feeding their (usually 5 or 10MHz) low phase noise crystal reference oscillator into a series of multiplier stages. Some stages might use a step recovery diode to create a rainbow of harmonics that they can then filter out a much higher order harmonic. Different manufacturers use slightly different designs, but they generally follow that architecture.

[@tauhidulkhandaker1068]

Very good. Loved It.

[@perdidoanonimo]

very nice, I really enjoy your videos!

[@IMSAIGuy]

thank you

[@____________________________.x]

I’m never going to do this, but I like knowing that I could, thanks

[@jyotirmoysarma1523]

Thank you it was very helpful

[@qwerty_____146]

Thank you, man. greattttttttttttttttttttt.

[@vidasvv]

Nice demo ! I think the "counter" function is standard on that series of HP SA's

[@IMSAIGuy]

thanks

[@charlessmith833]

So if I wanted a calibration source for my shortwave receiver across the band using harmonics it looks like I should generate a square wave from my crystal instead of a sine. Right? In other words a 100KC crystal changed to a square wave output would give me better harmonics for calibration across the dial? Please correct my thinking if I am wrong.

[@IMSAIGuy]

well yes. 100kc would give 200 300 400 .... but will drop off in amplitude so hope it reaches the frequencies you need.

[@mikaelkalki1129]

It is impossible to create an audible square wave as a perfectly vertical line in a wave form is physically impossible due to the continuous passage of time unless you can teleport air itself from point a to point B and back and forth

[@IMSAIGuy]

just curious. what is your age?

[@Master_zzz]

It's a nice trick to extract a harmonic from a square wave but also an expensive one. You are stuck with expensive LPF or BPF. I did it once building an 87.5MHz...108MHz FM DDS with a DDS only capable of outputting 12.5MHz max. Good luck building a BPF with steep rolloff.

[@nfachannel7928]

awesome, i want to understand more about harmonic signal and how to solve it, can you detail the filter circuit sir?

[@IMSAIGuy]

there are many many filter designs. Here is one: kzbin.info/www/bejne/foikaoCYgZqXpcU if you run a 7MHz square wave through that filter, it will remove all the higher harmonics and you will be left with a 7MHz sine wave. another one: kzbin.info/www/bejne/ZmWqk4qMm696kM0

[@jdmccorful]

Conditions this would be applied to?

[@IMSAIGuy]

I don't understand the question

[@lightbulbgonewild3205]

The harmonics: for example 100hz sine wave 300hz 500hz 700hz 900hz and so on...

[@IMSAIGuy]

yes, all the odd harmonics

[@azav8raa]

Great video, but you should fix the statement at 1:12 (kzbin.info/www/bejne/nmebmn1uf857fLM) about the power in the odd harmonics. They should fall off at 1/n. Questions: 1) which option in the 8591E includes the frequency counter? That is rather cool! 2) Can the 8591E be attached to a 10Mhz external reference like an GPSDO. Looking at a couple HP8591s on on e-bay.

[@amirb715]

frequency counter is under marker functions in almost all professional spectrum analyzers and no option is needed for that. 8591 can be fed with a 10MHz external reference.

[@IMSAIGuy]

I've checked my 8591 against my rubidium and it is below 1Hz error

[@azav8raa]

@@IMSAIGuy what options does your 8591 have?

[@IMSAIGuy]

@@azav8raa 21 and 140 HPIB Oven 10MHz 30Hz bandwidth

[@azav8raa]

@@IMSAIGuy Thanks. Option 140 would be very nice to have. From the User’s guide... Narrow Resolution Bandwidths and Precision Frequency Reference (Option 140) For HP 8591E, HP 8593E, HP 85943, HP 85953, and HP 85963 only. This option is a combination of Option 130 and Option 004. Option 140 provides additional narrow resolution bandwidths of 30 Hz, 100 Hz, and 300 Hz. These bandwidths improve sensitivity and allow you to resolve closely spaced signals. The option also includes an internal precision-frequency reference that improves stability and provides increased absolute frequency accuracy. The precision frequency reference makes the narrow resolution bandwidths more effective and easier to use.

[@eetech_fix2393]

You’re dealing with a square wave, an assortment of harmonics from its very nature. Your amplification and your capacitor arrangement just aggravate the system. If you look at the output on a spectrum analyzer, it’ll be obvious what you’re dealing with. Input a sine wave that’ll give you better results. Your scope is just triggering the predominant foreground harmonics. You’re missing what is there.

[@IMSAIGuy]

I would guess you didn't watch the video. I explain the harmonics of a square wave and also use a spectrum analyzer. I hope you have fun with the hobby and continue to learn

[@eetech_fix2393]

@@IMSAIGuy That’s amazing. I had two screens up, and I posted that comment to the wrong video. That comment was intended for another video experiencing a problem with a Square wave; your video was a perfect example of the problem they were having. Sorry about that

[@IMSAIGuy]

OK, carry on 😀

[@amirb715]

if the odd harmonics all had equal amplitude you would never get a square wave, in fact it would not be any physical wave at all (Fourier series will not even converge). So please study the basics before making educational video. The odd harmonics fall off as 1/n (n=1,3,5,7,...). The even harmonics appear if the duty cycle is not perfectly 50%. Also linear rising, falling edges also contribute to even harmonics a little

[@subramanianr7206]

Exactly. Thanks for pointing out. I have seen the harmonics in the output of a HF amplifier. A push-pull amplifier is practically immune to even harmonics if the duty cycle is 50% as it should be. And designing a harmonics filter depends on the very same phenomenon; otherwise it may not be possible by all practical means.