New Mixer Topology

I do not pretend that it is just a multiplying mixer, I proved it mathematically. But I stop insisting on that for the moment, you seem actually to be in a state of pigheadedness which makes you obviously inaccessible to rational and scientific arguments. That is not problematic and can happen to everybody from time to time. I'm sorry for that and thus I will remain quiet on this topic until your mind will be open again.

Thierry,

I have added some more notes and image to what I Posted: 3/2/2013 12:53:33 PM  to try to enlighten you about the operation, and particularly to clearly show that your:

"The fact that your output waveform is much "cleaner" is IMHO due to the fact that you just ***tadaa*** reinvented the switched capacitor low pass filter. When pulsed, the internal MOSFETs of the 4016 behave together with the loading capacitor like a simple RC low pass whose time constant depends on the duty cycle of the switching signal: tau = C * R = C * Ron * (Ton + Toff)/Ton which explains that your signal gets the more cleaner the shorter the duty cycle because this will increase the value of the virtual resistor."

Is utterly irrelevant to the way my circuit is operating and not in any way something that could possibly be happening in my circuit  - If you had looked at the component values you would have seen this. It is utterly obvious that the reason for the cleanness of the waveform has nothing to do with a resistor being PWM'd and producing a "switched capacitor filter" operation.. The reason is that the signal is being sampled and held (SAH).

I think it fair to say that you can utterly run rings 'round me when it comes to mathematics, and that I  run rings 'round you when it comes to electronics! - (I do think you are at an advantage though - you are far better at electronics than I could ever even start to be with mathematics - and you are getting better with electronics, and have the advantage of sound mathematics to assist you - Yeah.. You have the advantage!)

Your description of the switched capacitor function being responsible, and the idea that I would not have realised this had it been happening, would be insulting if it wasnt so laughable -- Just as if I was trying to explain how to transpose a simple formula to you would be more laughable than insulting! ;-)

All I can say to you now is that you seem actually to be in a state of pigheadedness which makes you obviously inaccessible to rational and scientific arguments. That is not problematic and can happen to everybody from time to time. I'm sorry for that and thus I will not bring this matter up again unless provoked. Perhaps the additional information given will open your mind  again.

Fred.

I never insisted on the switched capacitor RC thing, it was a thought, not more, and I did not come back on that because I was not sure. In the meantime I understood that this thinking model makes absolutely no sense in your context because the mentioned low pass filter effect works only when the switching frequency is much higher than the switched signal.

That's why I insisted (and still insist) only on the sampling=multiplication thing. A quick look on my mathematics shows that all my argumentation is only about that and I did not come back on the RC idea because there is even no need for it. As I demonstrated mathematically, the sampling (multiplying) with relatively short pulses alone gives already a clean signal. Thus I confirmed the principle of operation of your circuit and your results and proved at the same time that I was wrong about the RC theory.

Is it possible that you were so fixed on my RC statement that you did not perceive that all my following postings were about a totally different topic?

"Is it possible that you were so fixed on my RC statement that you did not perceive that all my following postings were about a totally different topic?" - Thierry

No, alas - I wish it was that simple.. In fact, I wish that either I could see that what I have been saying is wrong, or that you could see that what you are saying is wrong! - If I saw it, I would own / admit it.. As I said earlier, it actually makes no difference to me in practical terms, because whatever it is it does what I want!  ... Yeah, If I saw I was wrong it would be uncomfortable and embarrassing for a while, but I could handle it... Probably easier than I can handle the present situation which feels like unjust and dishonest critisism.

But I really have no more to say on the matter - I can see that applying 50ns pulses to one input of a multiplier and a waveform on the other input will result AFTER DEMODULATION in a replication of the waveform (within the constraints of the harmonic frequencies being low enough).. But a <50ns pulse would I believe produce (through multiplication) a waveform (after 4Q demodulation and filtering) with a tiny total amplitude compared to what I am getting. Apart from this, the practical aspects of generating consistant width narrow pulse and mixing this with any conventional low cost electronics....

However - even aside from the practicalities - I cannot see how, if the relevant coeficients are near to 1 (or even if they are close to any constant value) multiplication is really relevant.

As I see it, your position on this is saying that ALL sampling and heterodyning and mixing of any kind are all the same - that there is no difference between the sample playback electronics from a CD player, and the sampling in a sampling musical instrument, and the mixer in an EW!

And that because they are all "multiplying" (in accordance with your mathematical proof) they are all one and the same.. And that because they are all one and the same, there is nothing new in it and no merit in it and - effectively - that I am talking out my RS.

And as I have said, you may be right - at an extreme, hypothetical pedantic pure mathematics level..

But for all practical levels, IMO, it is nonsense... I know that if I went for a job interview (even to a company like Moog)  and was asked to explain the operation of a sampling card, and I went off on one about it being a multiplying mixer the same as the EW heterodyning front end, the interviewers get me out the door as fast as possible!

And if I was interviewing someone for a job and they came up with such nonsense, I would get them out the door as fast as possible, LOL ;-)

So perhaps we should leave it like this:

I agree that if the sampling sound card in my PC is, in fact, in practical terms, a multiplying mixer , Then I agree that the scheme I have presented here is a multiplying mixer.

You agree that if the sampling sound card in my PC is, in fact, in practical terms,  NOT a multiplying mixer, Then You agree that the scheme I have presented here is NOT a multiplying mixer.

Deal ?

I really dont wish to discuss this further - Let people decide for themselves based on the above "tests".. Because as I see it, if my scheme is a multiplying mixer, then my sound card and sampling keyboards and CD player and MP3 player are all multiplying mixers..

And they may be - in the world of purly theoretical mathematics.. But for every practical purpose, unless pushed to the limits at which problems occur, thinking about them as multiplying mixers and bundeling them together in the same group as analogue multipliers is, IMO, nonsense.

Fred.

"I am talking out my RS" - Fred

Now that is funny and I got a gut ache laughing so let me match your wit.

Metaphorically:  The big event today here in the center of town is this Chili-Cook-Off where people “demonstrate” their best concoctions. Each judge carefully tastes the chili from a purposely place stoneware bowl. Delighted they move to the next participant, the finest chef in town, she says she has a brilliant chili recipe never before experienced but it’s formula is folded up on a piece of paper placed inside her empty bowl. She does not win and walks away. She is a renowned restaurant owner and begins to whine that no one wants to share with her their own secret chili recipes, then she swears to never come back. She storms away calling everyone’s chili concoction inferior rubbish and all the participants ignorant. (-'

Christopher

I agree that if the sampling sound card in my PC is, in fact, in practical terms, a multiplying mixer , Then I agree that the scheme I have presented here is a multiplying mixer. (...) But for every practical purpose, unless pushed to the limits at which problems occur, thinking about them as multiplying mixers and bundeling them together in the same group as analogue multipliers is, IMO, nonsense. -Fred

Just take the anti-aliasing filters away and bring the sampling rate close to the input frequency. Your sound card will be a perfect ring modulator.

With kindest regards from Mr. Nyquist

Those who can not trust my words can find more information and the complete theory here:  https://en.wikipedia.org/wiki/Sampling_(signal_processing)

Short citation from the above link: The Whittaker–Shannon interpolation formula is mathematically equivalent to an ideal lowpass filter whose input is a sequence of Dirac delta functions that are modulated (multiplied) by the sample values. When the time interval between adjacent samples is a constant (T), the sequence of delta functions is called a Dirac comb. Mathematically, the modulated Dirac comb is equivalent to the product of the comb function with s(t).

Just take the anti-aliasing filters away and bring the sampling rate close to the input frequency. Your sound card will be a pure ring modulator. - Thierry

And the above doesnt come into the "unless pushed to the limits at which problems occur" catagory ?

Ring modulators and analogue multipliers / mixers (which is what ring modulators are) produce sum and difference frequencies - my system does not!  My system (when operating in the way I describe within the constraints I have given) produces only the difference frequency as a SAMPLED waveform.

The output is a waveform at the difference frequency stepped at the frequency of the sampling oscillator.. No Sum is produced.. (its not that the sum is produced and filtered away - ITS NOT PRODUCED!  BECAUSE IT IS NOT A MULTIPLIER! )

Christopher -

I am really sorry - I honestly never had a thought of you in my head when I wrote that.. Long before I joined TW I used RS as an abreviation on postings elsewhere.. It was quite difficult here at TW occassinally trying to remember not to use it here because it was a members name. I admit that since our enmity, I have deliberately used it perhaps a couple of times - but this was not one of those times..

Fred.

THIS IDEA AND THE SCHEMATICS ETC RELATED TO IT ARE UNSUITABLE FOR ANYONE WHO DOES NOT UNDERSTAND SAMPLE AND HOLD CIRCUITS, AND FOR ANYONE NOT  ABOVE BASIC ELECTRONIC HOBBYIST LEVEL.. YOU CANNOT EXPECT TO JUST BUILD THESE CIRCUITS AND HAVE THEM WORK WITHOUT UNDERSTANDING THEM AND DOING THE REQUIRED ADAPTATIONS TO FIT THE APPLICATION! << This warning was added on 3/29/2013 1:18:28 AM >>

From the above link: https://en.wikipedia.org/wiki/Sampling_(signal_processing)

A sampler is a subsystem or operation that extracts samples from a continuous signal.

That is what my system does. That is its primary, top level, most significant function.

Had enough..

Had enough.. - Fred

Then stay please here with your self-resticted viewing angle.

I'll return in my world where people around me are interested in enlarging their horizon, in thinking outside the box, in seeing old things under new viewing angles, in looking for underlying common principles, and so on.

I did not write much here in the last time. Now I remember why I stopped. I do not want to be criticized or even to apologize for the greater lines which I see and which others don't want or cannot see.

I hope you find enough public here to sell your old wine in new skins.

Good night

Bye

ILYA: ...The next idea...
FredM: ...I have a design which uses a heterodyning ALU...

I don't know about ALU (lets Fred to do it ;-) ) but the idea with ADC works perfectly: the internal sample-and-hold circuit of ADC transfers signal from RF to audio frequency wery nice. Below are the real waveforms at input and output. Horizontal scale is 2.5 and 100 us/div. 

A waveform from VPO is specialy distorted (positive and negative half-waves are different) to demonstrate that the shape retention is to be. I used the ARM-based CPU (STM32 family) with 12-bit ADC and DAC, both. A "heterodyning" freq is the sampling freq (about 200kHz), which is obtained from CPU clock (24 MHz). In other words, the "FPO" is internal. A microware just sends the samples from ADC to DAC. 


In general, there is nothing interesting here. Fun starts with DSP algorithms.