Thanks, man!

You get an E- for punctuation

Thank you for the encouragement! Welcome to the "family"! :-)

Yes, it was a high level overview of the mixer itself. I used a "ready made" mixer that exists inside the IC. There is actually several ways that this balanced mixer is implemented, so the discussion can get a bit complicated if we were to cover all of the options. Probably one of the most common implementations used today is the "Gilbert Cell" which consists of a "matrix" of matched transistors and a small number of resistors. I could have a video on this one implementation easily enough. I will add it to my list of videos to produce. In the meanwhile, to satisfy at least some of your curiosity, you can find it here: eng.libretexts.org/Bookshelves/Electrical_Engineering/Electronics/Microwave_and_RF_Design_IV%3A_Modules_(Steer)/06%3A_Mixer_and_Source_Modules/6.03%3A_Single-Ended_Balanced_and_Double_Balanced_Mixers and here: rfic.eecs.berkeley.edu/~niknejad/ee142_fa05lects/pdf/lect18.pdf Hope this helps.

Your comment got my curiosity up and just **HAD** to play a bit. I created a Gilbert Cell balanced mixer using LTSpice (FREE circuit simulation program). Here is a link to my model: drive.google.com/file/d/1CAEdY_aOMJJHIAjicFtPyRCJHNRR45zj/view?usp=sharing It is by no means a perfect design example. I was just thrashing and playing. It was fun and a nice precursor to when I make my video on the subject. Here's the download link for LTSpice: www.analog.com/en/design-center/design-tools-and-calculators/ltspice-simulator.html

You are very welcome! I am glad that you found it helpful! 🙂

Praise the Lord that it was helpful and ... WOW! General class at 12! I got my Novice at 15 and General at 16 back in 1971.🙂

@@electronicsfortheinquisitiveex she was actually 11, but she really worked hard for it.

@@halledwardb WOW! One smart cookie! 🙂

You are very welcome! 🙂

Thanks! 🙂

Interesting analogy

Thank you for the encouragement! :-)

You are very welcome! 🙂 I am glad that you found it beneficial.

@@electronicsfortheinquisitiveex I add: you also seem to be a good and selfless man, the ideal person with whom I would have liked to have a good glass of wine in the evening, after dinner, to chat about philosophical questions. Have a good luck, my friend, you and the people you care about.

@@RobertoPietrafesa Thank you Roberto. I think I would like that, too! 🙂

First, with regard to the observed harmonics. Yes, IF we had a perfectly linear modulator (mixer) and perfectly pure signals to mix, we would have simply F1 +/- F2. However, as you rightly point out (and I mention in the text contained in the video description), we do not have either of these. So, in a real world mixer and with real world signals, we do, indeed, have all of these harmonics to deal with. 😞 Now, with regard to the whole "wasting power" issue. My point was that we are putting power into either redundant "information" or something that does not actively carry information, not that SSB uses less power. If I am transmitting with 100 watts, that power is spread across the entire bandwidth of my transmitted signal. This would include the carrier, if I am using AM. However, if I am using SSB, then this 100 watts is "concentrated" only in the single sideband which contains everything I need. So, in terms of DSB vs SSB, the power density of SSB is twice that of DSB IF I am transmitting with 100 watts in both cases. So, yes, bandwidth conservation ... absolutely. Again, this means that all of my power goes into the information I am looking to communicate. I am not conserving power, but concentrating my power in a single sideband.

Unfortunately, YT does not allow its creators to simply "replace" and existing video or I would refactor this one adding these explanations to it. I can either totally delete the original video with all of its history and the like or I can add a followup video. Unfortunately, few people actually read the description for the additional explanations. So, I suppose that a followup video is in order. I can place a link in the existing video to the followup. I do not know when I will get to it, but I am going to put this in the video queue.

@@electronicsfortheinquisitiveex I do get it - sorry I didn't see that. What some people do is put any additional explanations/clarifications in the comments, and pin it to be the top comment.

@@BrightBlueJim That is certainly a good idea. It would get more visibility that way.

Thanks! 🙂

Thank you and you are very welcome! 🙂

Really good question ... made me revisit something. If you remember from the video I observed that the double sideband suppressed carrier signal was missing its even harmonic sideband components. I saw only the odd ones at 1 KHz, 3 KHz and so on. I questioned within myself if this was just because of my homebrew balanced modulator. So, I repeated the experiment with my IC-7610. I saw the same thing and then some. So, besides the fact of the phase distortion and narrower audio bandwidth that I mentioned in the video, there is the fact that we lose sideband content. In the case of the IC-7610, there was ONLY the 1KHz sideband whereas AM had 1, 2, 3 and 4 KHz sidebands. I saved the data from my IC-7610 experiment in a spreadsheet and put it into a ZIP file. Here is the link to that zip file: drive.google.com/file/d/1eNx9kYhcokmrtD3wYco1Z4nT0EYcaZDr/view?usp=sharing I hope this satisfies your curiosity.

Hey Benjamin ... I did a deep dive into this whole question and the spectral content of an AM signal and all. I even worked through a mathematical model of AM and did an FFT of the resulting time domain signal. It turns out that the **IDEAL** AM signal will have the carrier with a single pip on either side of the carrier (sidebands). The extra pips at 2xFm, 3xFm, etc represent distortion in the signal. So, it looks like we are left with the audio bandwidth at both the transmitting and receiving end as well as the phase distortion that I originally spoke of in the video.

@@electronicsfortheinquisitiveex Thanks for the insightful replies and crunching the numbers. I postulate in SSB the lack of fidelity is primarily caused by phase distortion and the audio bandwidth issue is negligible. One way to find out: What if we somehow restricted an incoming baseband audio signal to 3khz going to an amateur transceiver transmitting in AM. And compare this control group to SSB with the spectrum analyzer hooked up to the output of the receiver. No I'm no EE, so I can't back it up with all the math, but what I suspect causes the phase distortion is the reinsertion of the carrier at the receiver. Even if the carrier is reinserted at exactly the right frequency it is no longer in phase. One further reason I suspect it is the phase distortion is: doesn't commercial FM broadcast stereo employ double sideband suppressed carrier? What is the phase reference? I bet it is the pilot tone! Thanks again!

@@margaqrt FM broadcast is Frequency Modulation which, by definition, continuously transmits a carrier plus sidebands on either side of that carrier. I am not sure all of what makes up a Stereo FM Broadcast signal (never paid any attention to it), but it is still FM. The mathematical equation for FM is **FAR** more complex than AM. It almost hurts to think of it. LOL ;-)

Isn’t it also a matter of filter bandwith? On some radios you have the possibility to select the filter regardless of the type of modulation. By doing that effectively you can experience a narrow AM with the “closed” sound to it or a wide and bright SSB audio signal. Thanks for this interesting analysis supported by intuitive examples.

Thank you and welcome to the "family"! I am so glad that it was helpful.

Thank you! I am very glad that you found it helpful! 🙂

Thank you so much! I am so glad you appreciated the video. 🙂

You are so welcome! :-)

Actually, if you look closely, there are sidebands there for double sideband, suppressed carrier spectrum. The problem is that the line is light blue and hard to see.

You are very welcome on both counts! 🙂Admittedly, I honestly was at a loss for the Ainos reference until I went back to view the video to discover what you were referring to. This was a picture that I downloaded from PIXABAY as an illustration of oldness. Thanks to you, I learned something today! So, thank you! 🙂

You are very welcome, my friend! 🙂

You are very welcome! :-) I just added a link to a ZIP file which has an Excel spreadsheet in it comparing the spectral content of AM, LSB and USB using my IC-7610 in the video description. Here is the link for your convenience if you are interested: drive.google.com/file/d/1eNx9kYhcokmrtD3wYco1Z4nT0EYcaZDr/view?usp=sharing

Thanks! 🙂 Yup! There is only so deep one can go in a video this short. I **DO** plan on a video on the Gilbert Cell balanced mixer. It is actually very much in the works. But, admittedly, it will be a few weeks before I get to it.🤓

@@electronicsfortheinquisitiveex Keep up the good work ! Appreciated.

@@ernestb.2377 Thanks! 🙂

Welcome to the "family!"

Thank you! :-)

Gooooood question! Hang on for the answer, because to explain where the sidebands come in, we have to dive into the math to see it. Remember that I said that a modulator is a “multiplier?” Step 1: Our RF is sinusoidal. So we can write the following for the carrier in the time domain: v(t)=Ac*cos(ωc*t) Where Ac is the amplitude of the carrier, ωc is the frequency of the carrier in radians-per-second and t is the time. Step 2: Now we have our modulating signal. We can write a similar equation for the modulating signal: m(t)=M*cos(ωm*t) where M is the amplitude of the modulating signal, ωm is the frequency of the modulating signal in radians-per-second and t is time. Step 3: We modulate! The mathematical equation for Amplitude Modulation will be Vm(t)=Ac*cos(ωc*t)+(M*cos(ωm*t))*cos(ωc*t) Looking at this last portion a little bit, we have the amplitude of the modulating signal times the modulating signal itself times the carrier signal. Now, I pull out the CRC Math tables to find that cos(A) * cos(B) = ½*cos(A-B) + ½*cos(A+B) So, for this last bit of the modulation equation I will set A = ωc*t and B= ωm*t so we get: M*(½*cos(ωc*t - ωm*t) + ½*cos(ωc*t + ωm*t)) = M*(cos((ωc - ωm)*t) + cos((ωc + ωm)*t))/2 Notice the (ωc - ωm) and the (ωc + ωm)! The carrier frequency MINUS the modulating frequency and the carrier frequency PLUS the modulating frequency. The complete equation now looks like this: Vm(t)={Ac*cos(ωc*t)} + {M*(cos((ωc - ωm)*t) + cos((ωc + ωm)*t))/2} Vm(t)={Carrier} + {Sidebands} The first part of the equation is the carrier. The second part of the equation are the sidebands. You can try this in EXCEL (I just did for fun using 250KHz for a carrier and 1 KHz for the modulating signal) and, when you plot a graph of the results, you will see an amplitude modulated signal. At this point, I either blew more fog into the situation or I burned off some of it. Hope it is the latter of the two. 🙂

@@electronicsfortheinquisitiveex I'll come back to you on that :)

You and I, and many others, struggle with the same thing. It starts out with “here is the amplitude variation on the oscilloscope” and immediate jumps to “the carrier amplitude is constant and there are the sidebands” on the analyzer. I believe sidebands aren’t real, they only exist because of the analyzer math.

@@nohrtillman8734 I'd politely challenge your disbelief in sidebands being other than a mathematical curiousity by merely stating that analogue SSB radios exist. 😃

@@DennisSantos Pehaps it has been a 70 year old marketing ploy, and that SSB switch really doesn’t do anything. 🙂

Thanks! Now that is an interesting thought. You should give it a try and see how it works. :-)

@@electronicsfortheinquisitiveex Well I did not know that , but such modulation standard exist. One of them is ACSB (Amplitude-Compounded Single-Sideband Modulation)

Well, my friend, in this case the balanced mixer I used operates at 455 KHz, so the RF source has to be 455 KHz. It is a fixed frequency source. The signal then goes into a 455 KHz IF which feeds another mixer. This is where the variable frequency local oscillator comes into play. One the other side of the coin, the product detector. Again, this chip, when used as a product detector, is expecting 455 KHz. It uses a fixed frequency source and the I.F. coming from the I.F. chain at 455 KHz. Ahead of the I.F. chain is the mixer responsible for taking the R.F. and down shifting it to 455 KHz. The local oscillator feeding this mixer would have to be set to a frequency that is 455 KHz off of the receive frequency to do this. This mixer is followed by the I.F. chain which consists of amplifiers and filters so the anticipated artifacts of the mixer do not make it to the detector. Hope this helps

@@electronicsfortheinquisitiveex It does indeed help. Thank you. The only thing I'm curious about now is it sounds like the 455KHz must be some established standard. Otherwise it sounds like the source station could be modulated with 455 and if no standard then a receiving station could be using something else entirely. Or, am I missing something related to this xmt/rcv situation....? Thanks.

@@NickFrom1228 Yup, 455 KHz is a very standard I.F. frequency that has been in use for a very, very, VERY long time. It is not the only standard, however, but it is probably the oldest. It was chosen so as to avoid a bunch of artifacts associated with mixers that cause problems downstream. From the receive perspective, it really doesn't matter what the I.F. frequency of the transmit station was as long as we add in a carrier at our I.F. frequency at the receiving end.

@@electronicsfortheinquisitiveex Excellent. That clears it up nicely. Thank you, God bless and have a great weekend.

@@NickFrom1228 You have a great weekend, too! 🙂

Good question! Think of it this way ... the frequency of the audio that you hear is relative to the position of the carrier. When you change the frequency that you are listening to, you are changing the position of that carrier while the sidband signal has not moved. Let's say that you are listening to USB. If the carrier is positioned in its original location, then all of the sideband signal is above the carrier position. As you tune up in frequency, you are increasing the frequency of the injected carrier which is moving it closer to the unmoving sideband signal. Its *relative* position is closer and, thus, the audio frequency goes down. Hope this helps dispel the mystery. 🙂

Thank you very much - this has been bugging me. One follow-up if you don't mind - the "injected carrier" you reference is injected by the receiver, correct? @@electronicsfortheinquisitiveex

@@rick2194 Yes, that is correct. It is the one that is required in order to "make sense" of the SSB signal being received.

Yes, the glories of digital signal processing (DSP). They suck the digital samples of the SSB signal into a processor and digitally reproduce the original audio. I am not sure how they do that as I am *definitely NOT* a DSP guy, but they can do amazing things with sampled data.

@@electronicsfortheinquisitiveex oh ..... Thank you so much for clearing my doubt 😀👍🏻

Well ... I kinda did. I just didn't call it that (though I probably should have). I talked about how the carrier had to be added back in. This was the job of the Beat Frequency Oscillator. It was a bit difficult to use (I did that on my first receiver, a Hallicrafters SW-25), but it could be done.

If you have a perfect sine wave and your mixer/modulator does not contain any non-linear characteristics, then you will see one pip at the modulation frequency on either side of the carrier frequency (DSB).

@@electronicsfortheinquisitiveex Thank you for your answer, gentleman. I meant a pure signal, not a modulating signal over a carrier. In my case, again, will I see only one vertical line on the spectrum analyzer, correspondig to that sisgnal? (I'm trying to find out a book, but all books (signal & systems) have an heavy mathematical base which I would like to overlook, looking for the practical conclusions. I would be grateful to you if you can suggest me this kind of document/book, a sort of "signal & system for dummies", that can explain the real examples of modulated signals emitted by a radio and how we can see these on a time domain and frequency domain. Thank you very much indeed again. 73 de IZ7VHF.

@@RobertoPietrafesa No modulation of any kind - a single pip on the screen which is the signal in question. Here is an interesting read to parse out around the math: drive.google.com/file/d/1AvVF_RyaYbevQHYDLhYaxsXs8IWPQy9o/view?usp=sharing

@@electronicsfortheinquisitiveex Thank you Mr Ralph for this suggestion! I will read it deeply and soon. Clark Gable is to american actors as Ralph Gable is to youtube trainers! 🤩

@@RobertoPietrafesa There is also these four videos that might be helpful: ukposts.info/have/v-deo/rpx6fYR7gGab0Hk.html and ukposts.info/have/v-deo/jJGDbHeQfqCi12Q.html and ukposts.info/have/v-deo/qH1ki5ConXlntJc.html and, lastly ... ukposts.info/have/v-deo/n4CGpY2PfI16mmg.html

The additional sideband signals that we see are from harmonic distortion in the modulated signal. If we had a pure, perfect 1 KHz tone (without any harmonic distortion in it) and a perfect modulator, then we would only see the single 1 KHz sideband pip. But, alas, nothing is perfect. So we get this additional "stuff." The only way to see it is using a frequency domain device like a spectrum analyzer or an oscilloscope with FFT capability. Hope this helps dispel the fog a bit.

@@electronicsfortheinquisitiveex Thank you for the answer! I see now. So If I had a lot of distortion I would have a lot of other frequencies on spectrum analyser? Or it is the case also when I can hear my 27 MHz singal on my radio tuned to 2,7 MHz ? So it is like a bad copy of my oryginal singal in different frequencies? I am trying to understand this. I always discover something strange which doesn't follow the explanation from the book.. I know that the examplaes are ideal examples but still some authors can't explain simple things or maybe they can't ...

@@grzesiek1x Yes. The more distortion, the more frequencies you would see on the spectrum analyzer. You could have some Excel fun with this... An AM signal has the equation v(t) = (1+sin(2*pi*fmod*t))*cos(2*pi*fcarrier*t) where fmodulation is the frequency of the modulation, fcarrier is the frequency of the carrier, t is time in seconds. This assumes a sine wave modulating the carrier. Excel will do an FFT. You use this equation to generate the time domain signal, use Excel to do the FFT to get the frequency domain stuff.

@@electronicsfortheinquisitiveex Thanks again, I will try it :)

May I try to add some basic explanation for you ? It is some understanding that got to me recently, so I am glad to show it. Have you ever tuned an instrument with a tuning fork ? Lets say a guitar, and the snare to be tuned is slightly off. When you strike the snare and fork at the same time, you will hear 4 things: the frequency of the fork, the frequency of the snare, AND the difference between these two, AND the addition of the two, often less noticeable. When the fork is 880 Hz, and the snare is tuned as 882 Hz, you will also hear the 2 Hz difference as a wining extra sound. And there's also the total (addition) of the two basic tones, 880+882=1762 Hz, less noticeable than the 2 Hz. In this story, the 2 Hz and the 1762 are side bands, extra wave energy left and right of the two main frequencies of 880 and 882 Hz. By the same principle two extra energy bands form when you take a carrier, and ADD a modulation on it. The guitar example were low frequencies, and the difference was small (2 Hz), while the addition was large (1758 Hz) compared with the ground frequency of 880 Hz. But if you use an example of 10 MHz with a modulation of 1KHz, the subtraction is 9.999.000 Hz and the addition is 10.001.000 Hz. Two extra frequencies are formed, the side bands, with the carrier in the middle. When you would not use 1000 Hz modulation, but spoken word, a whole spectrum is formed as side bands close to the carrier. I think it is not possible to see these bands on a normal scope, because they linger to close to the carrier frequency. That is why they use a spectrum analyser, so you can see how the different frequency energies are distributed around the main frequency.

That is a good question. One which I have no answer for. Of course, FM isn't subject to interference (noise sources like static) like SSB is.

Your question is not dumb. Most of my experience was with 2m, so that informs my comments. The band covers 144- 148 MHz, which is a large piece of spectrum that has never been very crowded compared to the HF bands. Large portions of 2m have never been regularly utilized, so any transmission mode would not be an issue with respect to spectrum crowding. Since most communication on VHF and UHF bands is line of sight, NBFM produces a higher quality received signal in mobile and portable operation. Equipment, both transmit and receive, is relatively simple to implement (compared to SSB) and small vertical antennas work well when communication takes place through a repeater. On the other hand, DX is possible on 2m when certain atmospheric conditions are present or when bouncing signals off the moon or Aurora Borealis. In those circumstances, CW and to a limited extent SSB are superior. I got my ham license in the early 60s while I was in 9th grade. I was interested in the technical aspect of the hobby and not very keen on CW. With a Technician license, I could operate phone on 2m and that's where I started with a Heathkit Twoer (lunchbox) AM transceiver. It only put out about 1 watt and had a super regenerative receiver which wasn't very good (broad as a barn and deaf to all but the strongest signals). As time went on, I bought brand new Army surplus ARC5 receiver for $15 and built a 6CW4 nuvistor converter. Now I could hear a lot better, but nobody could hear me. Thanks to another local ham and great army surplus parts availability, I built a linear power amplifier from a donated 829B tube. I used the twoer to drive it and started increasing my reach to stations further away. Soon I noticed SSB activity and wanted to contact those stations as well. I found an article in CQ magazine that described how to build a 6m SSB transmitter. I scrounged enough parts to put it together, but needed a way to get it on 2 meters. QST published an article on how to build a 6m to 2m transmitting converter, so I built it. Soon I was on 2m SSB with about 60 watts from the 829B linear and was making contacts deep into Ohio, Ontario and western NY (I'm in Michigan). I was hearing stations running more power but they couldn't hear me. By now I was a senior in high school and when I graduated, my parents gave me a 60' tower as a graduation present. My dad and I erected it and put an 8 element beam on top. I still needed more power and started gathering parts to build a linear amp based on a pair of Eimac 4CX250Bs. I met a lot of great hams locally, and one gave me a pair of used tubes which made it possible for me to finish the amp. I worked 13 states on 2m SSB with some of them via the Aurora (point the antenna north and listen for ghostly CW signals or SSB that sound like whispers). College and other interests displaced my ham radio days and I gave all of my equipment away. After I left home, my dad sold the tower. I never got bitten by the 2m NBFM and repeater bug, but my senior design project in college was to design and build a 2M NBFM transceiver which sits on the shelf today. I think 2m NBFM killed a lot of the potential for 2m SSB, kind of like MTV killed the radio DJ!

I'm glad that you enjoyed the video. :-) Well, the reason why no one shows the difference between LSB and USB on the oscilloscope is because there is no difference between LSB and USB as viewed on the oscilloscope. Where the difference is observed in with a spectrum analyzer (frequency domain). Here we can see the LSB envelope below the carrier frequency and the USB envelope above the carrier frequency. In the time domain, there is no difference between the two.

@@electronicsfortheinquisitiveex Thanks.

@@cosmefulanito5933 🙂 You are welcome

When you listen adults speak on the old Peanuts tv specials, it always sounded to me like SSB without a beat frequency present. I always wondered if that’s what they did to create that distortion.

Playing devils advocate to jump start my brain into understanding this AM frequency shift thing. Oscilloscope = real time representation of a signal. Frequency Analyzer = math applied to the time domain. Does it stand to reason that sidebands only exist because of the Fourier math applied to the real time domain capture?

You are absolutely correct! :-) I discovered this fact AFTER I posted the video and added comments to this effect in the description. Thank you for pointing that out. Always learning we are!