Author Topic: Frequency Divider for older Oscilloscopes??  (Read 8519 times)

0 Members and 1 Guest are viewing this topic.

Online tggzzz

  • Super Contributor
  • ***
  • Posts: 21225
  • Country: gb
  • Numbers, not adjectives
    • Having fun doing more, with less
Re: Frequency Divider for older Oscilloscopes??
« Reply #25 on: July 18, 2021, 06:11:23 pm »
It strikes me that people in this conversation would benefit from thinking about how "Tayloe Mixers" work. Also see "Polyphase mixers" and "N-path filters".

https://www.eevblog.com/forum/rf-microwave/polyphase-or-n-path-mixer/msg3381802/#msg3381802 isn't a bad starting point, but there are probably better ones.
There are lies, damned lies, statistics - and ADC/DAC specs.
Glider pilot's aphorism: "there is no substitute for span". Retort: "There is a substitute: skill+imagination. But you can buy span".
Having fun doing more, with less
 

Offline bdunham7

  • Super Contributor
  • ***
  • Posts: 8175
  • Country: us
Re: Frequency Divider for older Oscilloscopes??
« Reply #26 on: July 18, 2021, 07:04:12 pm »
If you take a single sample per period and repeat the same thing for 10,000 periods, when you reconstitute the signal, it will be an exact copy of the original signal, but its frequency will be 10,000 times lower.  By doing that it is possible to observe a 100 Mhz signal as if it were a 10 Khz signal, something that any cheap oscilloscope can do.

You might want to draw a picture or something because something is missing here.

If you take 10,000 samples at exactly the same rate as the input signal, which is what you seem to be implyiing, you'll have 10,000 of exactly the same sample--a flat line.

Second, you are clearly not storing the samples anywhere, so how are you 'reconstituting the signal'?  I believe the systems 2N3055 refer to require storing the samples.

Third, are you implying that the fact that you take exactly 10,000 samples will result in an apparent frequency reduction of 10,000 times?  I don't see how that would work and I don't see how your system could manage that in any case--it is too simple with just those components.

The only way I can see something with that few components working is as some sort of mixer.  You wouldn't have a schematic or anything by chance?
A 3.5 digit 4.5 digit 5 digit 5.5 digit 6.5 digit 7.5 digit DMM is good enough for most people.
 
The following users thanked this post: Jorge Ginsberg

Online 2N3055

  • Super Contributor
  • ***
  • Posts: 7463
  • Country: hr
Re: Frequency Divider for older Oscilloscopes??
« Reply #27 on: July 18, 2021, 07:44:18 pm »
If you take a single sample per period and repeat the same thing for 10,000 periods, when you reconstitute the signal, it will be an exact copy of the original signal, but its frequency will be 10,000 times lower.  By doing that it is possible to observe a 100 Mhz signal as if it were a 10 Khz signal, something that any cheap oscilloscope can do.

You might want to draw a picture or something because something is missing here.

If you take 10,000 samples at exactly the same rate as the input signal, which is what you seem to be implyiing, you'll have 10,000 of exactly the same sample--a flat line.

Second, you are clearly not storing the samples anywhere, so how are you 'reconstituting the signal'?  I believe the systems 2N3055 refer to require storing the samples.

Third, are you implying that the fact that you take exactly 10,000 samples will result in an apparent frequency reduction of 10,000 times?  I don't see how that would work and I don't see how your system could manage that in any case--it is too simple with just those components.

The only way I can see something with that few components working is as some sort of mixer.  You wouldn't have a schematic or anything by chance?

It is called RIS (random interleaved sampling) and ETS (equivalent time sampling).

You need to have a trigger circuit with very low jitter.
You have a sample and hold circuit (S/H) that takes super short sample (let say samples single value for 100 psec) and keeps it for readout.
So you trigger, and take a 100 psec sample.  You then readout that value and read it with slow A/D converter. In meantime, millions of periods of signal passed by. But since signal is repetitive we don't care. After saving that first point, we rearm trigger, and wait for new one. When it fires, then magic hapens: we sample signal with a 100 psec delay from trigger. Sample and save that. Third time, we take sample at 200 psec delay. That way we reconstruct a waveform from many many different trigger events. Equivalent sampling rate would be 10 GS/sec.

With RIS (random interleaved sampling) we also trigger, but it is upside down: we sample at random intervals, not sequentially from the trigger. But we do keep triggering, and measure time from trigger and sample so we know where to position the point in time domain. It is funny to look at how it assembles the waveform by plotting random  dots left and right until it starts to resemble the waveform...

That is how analog sampler scopes worked, with different details in implementation of S/H circuit, and actual timing and trigger circuits. Some samplers didn't even have real trigger circuit, but you had to provide trigger yourself, already conditioned...
"Just hard work is not enough - it must be applied sensibly."
Dr. Richard W. Hamming
 

Offline bdunham7

  • Super Contributor
  • ***
  • Posts: 8175
  • Country: us
Re: Frequency Divider for older Oscilloscopes??
« Reply #28 on: July 18, 2021, 08:28:34 pm »
With RIS (random interleaved sampling) we also trigger, but it is upside down: we sample at random intervals, not sequentially from the trigger. But we do keep triggering, and measure time from trigger and sample so we know where to position the point in time domain. It is funny to look at how it assembles the waveform by plotting random  dots left and right until it starts to resemble the waveform...

That is how analog sampler scopes worked, with different details in implementation of S/H circuit, and actual timing and trigger circuits. Some samplers didn't even have real trigger circuit, but you had to provide trigger yourself, already conditioned...

I understand ETS and sampling scopes, I even have several still.  I don't see how what Jorge is describing could possibly do any of those things.
A 3.5 digit 4.5 digit 5 digit 5.5 digit 6.5 digit 7.5 digit DMM is good enough for most people.
 

Offline Jorge Ginsberg

  • Contributor
  • Posts: 20
  • Country: ar
Re: Frequency Divider for older Oscilloscopes??
« Reply #29 on: July 18, 2021, 09:26:56 pm »
A picture is worth a thousand words, the Chinese used to say.
I would like to have a simple program with which I can draw sinewaves and other symbols to be able to explain in a single drawing what I find so difficult to explain with words...
Does anyone know of a light program that would allow me to do that?
 

Online 2N3055

  • Super Contributor
  • ***
  • Posts: 7463
  • Country: hr
Re: Frequency Divider for older Oscilloscopes??
« Reply #30 on: July 18, 2021, 09:55:33 pm »
With RIS (random interleaved sampling) we also trigger, but it is upside down: we sample at random intervals, not sequentially from the trigger. But we do keep triggering, and measure time from trigger and sample so we know where to position the point in time domain. It is funny to look at how it assembles the waveform by plotting random  dots left and right until it starts to resemble the waveform...

That is how analog sampler scopes worked, with different details in implementation of S/H circuit, and actual timing and trigger circuits. Some samplers didn't even have real trigger circuit, but you had to provide trigger yourself, already conditioned...

I understand ETS and sampling scopes, I even have several still.  I don't see how what Jorge is describing could possibly do any of those things.

Funny thing is that sometimes non native English speakers understand better other non native English speakers because to us it is a logical process, not natural simple recognition. We go by logic of the said words and not by what it should be in good English.  Trust me, he's is talking about sequential sampling principle (ETS style).  And if you do 10000 points per second, you don't even need A/D and storage. You simply show it on slow CRT scope screen where phosphor persistence will act as a storage for some miliseconds, enough to build full screen.
"Just hard work is not enough - it must be applied sensibly."
Dr. Richard W. Hamming
 

Offline Jorge Ginsberg

  • Contributor
  • Posts: 20
  • Country: ar
Re: Frequency Divider for older Oscilloscopes??
« Reply #31 on: July 18, 2021, 10:05:29 pm »
I made this drawing using paint.  It's not a good drawing, but just by looking at it you will understand exactly how I was able to get a 100 Mhz signal to transform into a 10 Khz signal.  If the samples are thousands per cycle, the reconstituted waveform is a perfect copy of the sampled signal.  See the attached png file.
 

Offline bdunham7

  • Super Contributor
  • ***
  • Posts: 8175
  • Country: us
Re: Frequency Divider for older Oscilloscopes??
« Reply #32 on: July 18, 2021, 10:52:04 pm »
Funny thing is that sometimes non native English speakers understand better other non native English speakers because to us it is a logical process, not natural simple recognition. We go by logic of the said words and not by what it should be in good English.  Trust me, he's is talking about sequential sampling principle (ETS style).  And if you do 10000 points per second, you don't even need A/D and storage. You simply show it on slow CRT scope screen where phosphor persistence will act as a storage for some miliseconds, enough to build full screen.

Sometimes by approaching things from different angles, two people can describe the same phenomenon in different ways.  If you look at his picture and read my description, I think they match exactly.  How would you describe that other than as an alias?  It's only going to work if the sampling frequency is offset from the signal frequency slightly, and by a known amount if you want to have a predictable display.

I'm not sure what you mean when you say that you can just put it on the screen at 10,000 samples per second.  ETS also has a way of knowing the timing of each sample so that it knows where (horizontally) to display it.  Can you show me an example where this would work otherwise?  Without storing the samples, to put them up at the right position would require the sweep rate to be synced to the input frequency.  Unless I'm missing something, which is always possible, I think the method described here is some sort of alias-based downmixing and can't work for arbitrary frequencies unless the sampling rate is somehow automatically tuned with a fixed offset.

Now if the circuit is complex enough that it can trigger on the 100MHz signal and then produce a ramp to trigger the sample point at calibrated and increasing time offsets, then it might be able to accomplish the task without depending on any particular characteristic of the input other than repeatability.  Then you'd have a phosphor-based storage sampling scope like you describe.  I don't see that what he describes could accomplish all that--it would require a bit more complexity than he describes.
« Last Edit: July 18, 2021, 11:01:04 pm by bdunham7 »
A 3.5 digit 4.5 digit 5 digit 5.5 digit 6.5 digit 7.5 digit DMM is good enough for most people.
 

Offline bdunham7

  • Super Contributor
  • ***
  • Posts: 8175
  • Country: us
Re: Frequency Divider for older Oscilloscopes??
« Reply #33 on: July 18, 2021, 10:54:58 pm »
I made this drawing using paint.  It's not a good drawing, but just by looking at it you will understand exactly how I was able to get a 100 Mhz signal to transform into a 10 Khz signal.  If the samples are thousands per cycle, the reconstituted waveform is a perfect copy of the sampled signal.  See the attached png file.

OK, but it seems clear that your sampling rate is just slightly slower than the input signal, not equal to it, right?  How does the circuit determine and set the sampling rate?
A 3.5 digit 4.5 digit 5 digit 5.5 digit 6.5 digit 7.5 digit DMM is good enough for most people.
 

Offline Ian.M

  • Super Contributor
  • ***
  • Posts: 13216
Re: Frequency Divider for older Oscilloscopes??
« Reply #34 on: July 18, 2021, 11:05:43 pm »
I agree with Bdunham.  To down-convert a 100.000MHz signal to 10KHz, you need to sample at 99.990MHz, not at 100.000MHz.   If you want to automate this so that the output frequency is always 1/10000 of the input frequency, at the very least you'll need a good VCO for the sampling clock that covers the input range you are interested in, and to PLL lock it to 9999/10000 of the input frequency (which is a whole other can of worms if the input is not a nice quasi-sinusoid).

The requirements for the sample and hold are also rather stringent - it may well be easier to implement as a pair of carefully matched track and holds that alternate acquiring the signal, and a fast analog switch to select the one currently holding.
« Last Edit: July 18, 2021, 11:07:19 pm by Ian.M »
 

Offline bdunham7

  • Super Contributor
  • ***
  • Posts: 8175
  • Country: us
Re: Frequency Divider for older Oscilloscopes??
« Reply #35 on: July 18, 2021, 11:16:34 pm »
Just to be clear, I'm not claiming Jorge is wrong or that what he says he made didn't exist, just that I don't understand how it would work very well given the description.  I hope he is patient enough to continue the conversation and isn't put off by language barriers.  He mentioned some TTL logic, so perhaps something like a 7474? which would imply some trigger logic.  Making a workable trigger, ramp and sample/hold circuit that operates into the UHF region does not seem a trivial task.
A 3.5 digit 4.5 digit 5 digit 5.5 digit 6.5 digit 7.5 digit DMM is good enough for most people.
 

Offline Jorge Ginsberg

  • Contributor
  • Posts: 20
  • Country: ar
Re: Frequency Divider for older Oscilloscopes??
« Reply #36 on: July 18, 2021, 11:21:10 pm »
Quote
I agree with Bdunham.  To down-convert a 100.000MHz signal to 10KHz, you need to sample at 99.990MHz, not at 100.000MHz.   If you want to automate this so that the output frequency is always 1/10000 of the input frequency, at the very least you'll need a good VCO for the sampling clock that covers the input range you are interested in, and to PLL lock it to 9999/10000 of the input frequency (which is a whole other can of worms if the input is not a nice quasi-sinusoid).

The requirements for the sample and hold are also rather stringent - it may well be easier to implement as a pair of carefully matched track and holds that alternate acquiring the signal, and a fast analog switch to select the one currently holding.

No...no...no... and  NO.

Quote
OK, but it seems clear that your sampling rate is just slightly slower than the input signal, not equal to it, right?  How does the circuit determine and set the sampling rate?   

No...no...no... and  NO.

What changes is not the frequency of sampling.  What changes is the TIME at which the sampling occurs.  In the first cycle, (for example) it occurs 1 picosecond after the wave starts (zero crossing from bottom to top).  The second sample is taken 2 picoseconds after the zero crossing, from bottom to top.
The third sample is taken 3 picoseconds after the zero crossing, from bottom to top.
The fourth sample is taken 4 picoseconds after the zero crossing, from bottom to top.
The fifth sample is taken 5 picoseconds after the zero crossing, from bottom to top.
And so on until the cycle is complete.
And if you use 10000 samples to reconstitute the original waveform, and you have taken only one sample for each cycle, then the reconstituted signal will have a frequency 10000 times lower and you will be able to observe it on a cheap oscilloscope.
The sampled voltage is charged into a capacitor.  And on that capacitor a perfect sine wave is obtained, and a faithful copy of the sampled signal.
No A/D converters are needed.  No need for counters.  No need for memories.  The method and the circuit are much simpler than you suppose.  But you must use your wits to figure it out.  That's the beauty of electronics design: the ingenuity you use to solve a complicated problem with a simple solution.

The most remarkable thing about all this talk is that
GlennSprigg , the original author of the query, has not even shown up here......
« Last Edit: July 18, 2021, 11:39:03 pm by Jorge Ginsberg »
 

Offline bdunham7

  • Super Contributor
  • ***
  • Posts: 8175
  • Country: us
Re: Frequency Divider for older Oscilloscopes??
« Reply #37 on: July 19, 2021, 01:29:10 am »
The method and the circuit are much simpler than you suppose.  But you must use your wits to figure it out.  That's the beauty of electronics design: the ingenuity you use to solve a complicated problem with a simple solution.

OK, let's reason through it so at least we agree on the steps needed.  Then I'll think about how hard it would be to build.

Quote
What changes is not the frequency of sampling.  What changes is the TIME at which the sampling occurs.  In the first cycle, (for example) it occurs 1 picosecond after the wave starts (zero crossing from bottom to top).  The second sample is taken 2 picoseconds after the zero crossing, from bottom to top.

In many things there are different ways of describing the same thing.  Often two very descriptions will boil down to the same thing.  In this case, at a given signal (say our 100MHz) sampling as you describe will result in a sampling frequency just under the signal frequency, just by a different method.  In this case the method is important, but the result may be considered in different ways.

Quote
And if you use 10000 samples to reconstitute the original waveform, and you have taken only one sample for each cycle, then the reconstituted signal will have a frequency 10000 times lower and you will be able to observe it on a cheap oscilloscope.  The sampled voltage is charged into a capacitor.  And on that capacitor a perfect sine wave is obtained, and a faithful copy of the sampled signal.
No A/D converters are needed.  No need for counters.  No need for memories. 

OK, so to the meat of the problem.  Someone correct me if I err.  These are the issues, more or less in order.  Assume a 100MHz input signal to be displayed at 10kHz.

1.  You need a way to trigger on the zero crossing (or anywhere, I suppose) of the input signal.
2.  You are sampling at 100MSa/s using an offset that cycles at 10kHz and has a range sufficient to cover one period.
3.  If you want to display this as single period 10kHz signal on a scope, the scope should have a sweep time of 100 microseconds.
4.  You need a method of offsetting samples of the input frequency from 0 to 10 nanoseconds, corresponding to one period.
5.  To display the samples on the scope, they have to be correlated so that their positions in real time on the scope are proportional to the sample offset time.  So the +5nS sample must appear at the 50uS position.
6.  Since there is no storage, the samples have to occur in real time with the two values correlated.  Thus you need a circuit that 'knows' where it is on the 100uS scale and offsets the sample by the corresponding amount.

So you need a trigger that operates on the 100MHz input signal, a 10kHz ramp signal followed by an offset delay circuit that corresponds to the ramp and then operates the sample circuit.  You can trigger the display scope on the 10kHz ramp or on the sampled output.  Anything else?

Quote
The most remarkable thing about all this talk is that
GlennSprigg , the original author of the query, has not even shown up here......

Time zones.  We're all over the world and people pop in and out.
A 3.5 digit 4.5 digit 5 digit 5.5 digit 6.5 digit 7.5 digit DMM is good enough for most people.
 

Offline Jorge Ginsberg

  • Contributor
  • Posts: 20
  • Country: ar
Re: Frequency Divider for older Oscilloscopes??
« Reply #38 on: July 19, 2021, 02:55:15 am »
Quote
OK, so to the meat of the problem.  Someone correct me if I err.  These are the issues, more or less in order.  Assume a 100MHz input signal to be displayed at 10kHz.
1.  You need a way to trigger on the zero crossing (or anywhere, I suppose) of the input signal.
That's right. 
And any oscilloscope has a trigger circuit.  There is no difficulty in that.
Quote
2.  You are sampling at 100MSa/s using an offset that cycles at 10kHz and has a range sufficient to cover one period.
I'm afraid I don't understand what you're saying
Quote
3.  If you want to display this as single period 10kHz signal on a scope, the scope should have a sweep time of 100 microseconds
.
Once you have the reconstituted signal you send it to the oscilloscope of your choice.  To the oscilloscope, that reconstituted signal is just like any other signal it receives.  The oscilloscope is not obliged to know that you are feeding it a signal that was the product of another sampled signal...  In fact, you are doing all this to be able to see VHF signals with your cheap oscilloscope...
Quote
4.  You need a method of offsetting samples of the input frequency from 0 to 10 nanoseconds, corresponding to one period.
from 0 to 10 ns... from 0 to 5 ns... from 0 to 30 ns... Whatever you need.  This is varied with a potentiometer, preferably a 10 turns one.
Quote
5.  To display the samples on the scope, they have to be correlated so that their positions in real time on the scope are proportional to the sample offset time.  So the +5nS sample must appear at the 50uS position.
You are not feeding any sampled signal to the oscilloscope.  Your oscilloscope is receiving an already reconstituted signal and your oscilloscope doesn't care where that signal comes from nor does it care how it was created.
Quote
6.  Since there is no storage, the samples have to occur in real time with the two values correlated.  Thus you need a circuit that 'knows' where it is on the 100uS scale and offsets the sample by the corresponding amount.
No correlation with anything is necessary.  The circuit simply takes the samples one after the other, period after period, and a capacitor takes care of integrating everything and reconstituting the signal. 
What you must find out now is how to achieve that each sample is displaced with respect to the previous one and that this time is not modified, because if it is modified, the waveform will change and that is what we do not want to happen.
Quote
So you need a trigger that operates on the 100MHz input signal, a 10kHz ramp signal followed by an offset delay circuit that corresponds to the ramp and then operates the sample circuit.  You can trigger the display scope on the 10kHz ramp or on the sampled output.  Anything else?
No. Nothing else.
Put your ingenuity to work and see if you can figure out how to achieve all that in a simple way.
Enjoy the challenge !
 

Offline bdunham7

  • Super Contributor
  • ***
  • Posts: 8175
  • Country: us
Re: Frequency Divider for older Oscilloscopes??
« Reply #39 on: July 19, 2021, 03:54:59 am »

And any oscilloscope has a trigger circuit.  There is no difficulty in that.

No correlation with anything is necessary.
 
Put your ingenuity to work and see if you can figure out how to achieve all that in a simple way.


OK, I can see generally what is needed.  Making a reliable trigger may not be an insurmountable challenge, but picosecond jitter.... I suppose it doesn't have to be that good to nominally work.

Correlation will occur whether explicitly designed for or just an implicit result of things 'boiling down' to what they actually are.  The oscilloscope timing just serves conceptually as a framework, I understand the circuit doesn't depend on it in any way. 

I'll let GlennSprigg pick this up if he is still interested.
A 3.5 digit 4.5 digit 5 digit 5.5 digit 6.5 digit 7.5 digit DMM is good enough for most people.
 

Online tggzzz

  • Super Contributor
  • ***
  • Posts: 21225
  • Country: gb
  • Numbers, not adjectives
    • Having fun doing more, with less
Re: Frequency Divider for older Oscilloscopes??
« Reply #40 on: July 19, 2021, 08:02:15 am »
Here's a simple question that I have asked interviewees in a job interview...

"I want to observe a signal that is between 990MHz and 1000MHz. What is the minimum sampling frequency that I need to use?".

Those answering 2GS/s faced an uphill struggle.

(And I'm not surprised that the OP hasn't had the courtesy to respond)
There are lies, damned lies, statistics - and ADC/DAC specs.
Glider pilot's aphorism: "there is no substitute for span". Retort: "There is a substitute: skill+imagination. But you can buy span".
Having fun doing more, with less
 

Online tggzzz

  • Super Contributor
  • ***
  • Posts: 21225
  • Country: gb
  • Numbers, not adjectives
    • Having fun doing more, with less
Re: Frequency Divider for older Oscilloscopes??
« Reply #41 on: July 19, 2021, 08:22:19 am »

And any oscilloscope has a trigger circuit.  There is no difficulty in that.

No correlation with anything is necessary.
 
Put your ingenuity to work and see if you can figure out how to achieve all that in a simple way.


OK, I can see generally what is needed.

I have a 50 year old scope (battery powered, can be operated in driving rain and stored underwater :) ). It displays signals up to 4GHz with, IIRC, 37kS/s. It contains transistors, 741s and a few SSI logic gates (dual JK flip flops and nand gates), but no memory or anything like that. The full service manual contains schematics and circuit descriptions.

Watching the display twinkle is hypnotic, but I can't capture the effect in a video



Or see the attachment for another example circuit, from http://www.redrok.com/Circuits_1GHz-samplig-Oscilloscope-Front-End.pdf
« Last Edit: July 19, 2021, 08:24:01 am by tggzzz »
There are lies, damned lies, statistics - and ADC/DAC specs.
Glider pilot's aphorism: "there is no substitute for span". Retort: "There is a substitute: skill+imagination. But you can buy span".
Having fun doing more, with less
 

Offline Jorge Ginsberg

  • Contributor
  • Posts: 20
  • Country: ar
Re: Frequency Divider for older Oscilloscopes??
« Reply #42 on: July 19, 2021, 10:46:01 am »
Quote

OK, I can see generally what is needed.  Making a reliable trigger may not be an insurmountable challenge, but picosecond jitter.... I suppose it doesn't have to be that good to nominally work.

If you use semiconductors for UHF, if you use very stable voltages, and if you use constant temperatures, the jitter and drifts are so low that you hardly notice it.
 

Offline Jorge Ginsberg

  • Contributor
  • Posts: 20
  • Country: ar
Re: Frequency Divider for older Oscilloscopes??
« Reply #43 on: July 19, 2021, 11:18:06 am »
Here's a simple question that I have asked interviewees in a job interview...

"I want to observe a signal that is between 990MHz and 1000MHz. What is the minimum sampling frequency that I need to use?".

Those answering 2GS/s faced an uphill struggle.

(And I'm not surprised that the OP hasn't had the courtesy to respond)

Dear TGGZZZ... what kind of waves do you expect to find between 900 Mhz and 1000 Mhz? do you expect to find square waves? do you expect to find saw teeth?
When you work in the high UHF zone and enter the microwave zone, ALL waves are already sine waves.
An oscilloscope is useless at those frequencies.
At very high frequencies, a resistor can behave like a coil and a capacitor can behave like a delay line. Transistors already behave differently.  The "h" parameters are no longer used in the design, but the "s" parameters are used.
At very high frequencies all the electronics change.
Intel uses UHF and microwave engineers for the development of its chips.  Imagine a microprocessor running at 3.6 Ghz.....
A good technician or engineer in audio or industrial electronics becomes a perfect ignoramus when working above 1 Ghz.
It is true that to observe correctly a 100 Mhz square wave you would need a 1000 Mhz oscilloscope, otherwise the square wave would be very distorted in its rising and falling edges; but it is never necessary to "look" at 1000 Mhz waves. In the vast majority of cases it is enough to know if that wave exists and what is his amplitude. For example, people who buy the Tektronix MDO3000 oscilloscope, whose bandwidth is 1Ghz, do not do it to see 1 Ghz waves but to be able to see correctly the 100 Mhz waves, because when you look at a 1 Ghz wave, the only thing you see is a sinewave...
Those of us who work on UHF or microwave radio links only use spectrum analyzers.  By looking at the spectrum you know if you have just a sine wave or if you have a more complex wave.  In that frequency range it doesn't matter the shape of the waves, the only thing that matters is their harmonic content.
 

Offline Jorge Ginsberg

  • Contributor
  • Posts: 20
  • Country: ar
Re: Frequency Divider for older Oscilloscopes??
« Reply #44 on: July 19, 2021, 11:34:39 am »
The good thing about all this talk is that it doesn't end in a production.
I say again what I said before: nowadays none of this is worth it. You can buy a good 100 Mhz oscilloscope for 400 dollars or even less.
The appeal of all this talk is that we exercise our minds and find out how this device can be created analog, cheap and simple.
Think of the human ingenuity displayed throughout history.
In 1879, Albert A. Michelson measured the speed of light.  At that time, laser beams did not exist. There were no atomic clocks.
Michelson measured the speed of light using only mirrors and a rotating disk. And he determined the speed of light with an error of only 200 miles/hour !! AWESOME ...
In 1653 the astronomer Christiaan Huygens determined the distance between the earth and the sun. What technology existed at that time to achieve such a feat?
And more than 2500 years ago, the Babylonians already knew that the Earth was round and even calculated its diameter.
And all this was achieved thanks to human ingenuity.

That is why I wanted to tell you here that more than 30 years ago I built a device that allowed me to see UHF signals on a modest oscilloscope with a bandwidth of only 10 MHz.  And I will give you hints about how I did it because the truly beautiful thing about electronics is the way it sharpens our ingenuity.

If GlennSprigg was expecting to receive a ready-made design and simply copy it... I'm sorry to disappoint him.
 

Offline bdunham7

  • Super Contributor
  • ***
  • Posts: 8175
  • Country: us
Re: Frequency Divider for older Oscilloscopes??
« Reply #45 on: July 19, 2021, 01:55:55 pm »
"I want to observe a signal that is between 990MHz and 1000MHz. What is the minimum sampling frequency that I need to use?".

So I might respond by asking you to define the term "observe" and also ask whether  "between 990 and 1000MHz" means the entire spectrum of the signal falls between those bounds.  Would you respond or send me packing?  :)   Obviously sampling scopes can display anything that they can get a stable trigger on with an arbitrarily low sample rate if you are patient, but that doesn't mean they are cheap or easy to build.

what kind of waves do you expect to find between 900 Mhz and 1000 Mhz? do you expect to find square waves? do you expect to find saw teeth?

A lot can happen in a nanosecond these days....





A 3.5 digit 4.5 digit 5 digit 5.5 digit 6.5 digit 7.5 digit DMM is good enough for most people.
 

Offline Jorge Ginsberg

  • Contributor
  • Posts: 20
  • Country: ar
Re: Frequency Divider for older Oscilloscopes??
« Reply #46 on: July 19, 2021, 02:48:49 pm »
Quote
A lot can happen in a nanosecond these days....

Yes... but are you sure that what you are seeing is exactly what is happening?
When you measure signals in the microwave region you can never believe the waveform you are seeing.
Cables... connectors... even reflections of possible waves in the cabinet or shields ... all can alter the signal you are looking at.
When you open up a UHF circuit that is tucked inside a shield and try to observe some waveform, you are altering everything.
believe me... above 500 or 1 Ghz you work with spectrum analyzers, not oscilloscopes. 
In the most extreme case, it might be necessary to observe some trigger pulse that lasts several nanoseconds... but there you are not talking about frequency but time.  That trigger pulse could be repeated at the rate of one pulse every millisecond.  And someone could say "the frequency of the pulse is 1 kHz".  Nothing more wrong.  NEVER confuse times with frequencies when it comes to laboratory measurements.
 

Online tggzzz

  • Super Contributor
  • ***
  • Posts: 21225
  • Country: gb
  • Numbers, not adjectives
    • Having fun doing more, with less
Re: Frequency Divider for older Oscilloscopes??
« Reply #47 on: July 19, 2021, 03:03:00 pm »
Here's a simple question that I have asked interviewees in a job interview...

"I want to observe a signal that is between 990MHz and 1000MHz. What is the minimum sampling frequency that I need to use?".

Those answering 2GS/s faced an uphill struggle.

(And I'm not surprised that the OP hasn't had the courtesy to respond)

Dear TGGZZZ... what kind of waves do you expect to find between 900 Mhz and 1000 Mhz? do you expect to find square waves? do you expect to find saw teeth?

It is irrelevant, but you can consider a modulated signal.

Quote
When you work in the high UHF zone and enter the microwave zone, ALL waves are already sine waves.

Er, that's true in any frequency band.

If you don't like 900/1000MHz, divide it by 1000 or 1000000 and the question is equally valid.

Quote
An oscilloscope is useless at those frequencies.

There are 110GHz real time scopes commercially available. The Infiniim XUR series starts at 5GHz and goes to 110GHz albeit with only a 5bin ENOB. I can't afford one.

Now, how do you think Keysight implements it? With a >220GS/s 5 bit ADC? I doubt it (but haven't bothered to research the implementation), and I suspect the reason is a direct consequence of correctly answering my question.

Whether a scope is the right tool is a separate issue; frequently it isn't.

Quote
At very high frequencies, a resistor can behave like a coil and a capacitor can behave like a delay line. Transistors already behave differently.  The "h" parameters are no longer used in the design, but the "s" parameters are used.
At very high frequencies all the electronics change.
Intel uses UHF and microwave engineers for the development of its chips.  Imagine a microprocessor running at 3.6 Ghz.....
A good technician or engineer in audio or industrial electronics becomes a perfect ignoramus when working above 1 Ghz.

True but irrelevant to the point.

Quote
It is true that to observe correctly a 100 Mhz square wave you would need a 1000 Mhz oscilloscope, otherwise the square wave would be very distorted in its rising and falling edges; but it is never necessary to "look" at 1000 Mhz waves. In the vast majority of cases it is enough to know if that wave exists and what is his amplitude. For example, people who buy the Tektronix MDO3000 oscilloscope, whose bandwidth is 1Ghz, do not do it to see 1 Ghz waves but to be able to see correctly the 100 Mhz waves, because when you look at a 1 Ghz wave, the only thing you see is a sinewave...
Those of us who work on UHF or microwave radio links only use spectrum analyzers.  By looking at the spectrum you know if you have just a sine wave or if you have a more complex wave.  In that frequency range it doesn't matter the shape of the waves, the only thing that matters is their harmonic content.

My question is important for far more than scopes.
« Last Edit: July 19, 2021, 03:14:33 pm by tggzzz »
There are lies, damned lies, statistics - and ADC/DAC specs.
Glider pilot's aphorism: "there is no substitute for span". Retort: "There is a substitute: skill+imagination. But you can buy span".
Having fun doing more, with less
 

Online tggzzz

  • Super Contributor
  • ***
  • Posts: 21225
  • Country: gb
  • Numbers, not adjectives
    • Having fun doing more, with less
Re: Frequency Divider for older Oscilloscopes??
« Reply #48 on: July 19, 2021, 03:08:13 pm »
"I want to observe a signal that is between 990MHz and 1000MHz. What is the minimum sampling frequency that I need to use?".

So I might respond by asking you to define the term "observe" and also ask whether  "between 990 and 1000MHz" means the entire spectrum of the signal falls between those bounds.  Would you respond or send me packing?  :)   

No, I wouldn't send  you packing. Nothing special about the term "observe", use "see", "visualise", "measure" etc. Yes, all the energy does lie within those bounds; that's the key point :)

Quote
Obviously sampling scopes can display anything that they can get a stable trigger on with an arbitrarily low sample rate if you are patient, but that doesn't mean they are cheap or easy to build.

True if the signal is repetitive, but for a non-repetitive signal you have to capture it all on the fly (i.e. realtime).
There are lies, damned lies, statistics - and ADC/DAC specs.
Glider pilot's aphorism: "there is no substitute for span". Retort: "There is a substitute: skill+imagination. But you can buy span".
Having fun doing more, with less
 

Online tggzzz

  • Super Contributor
  • ***
  • Posts: 21225
  • Country: gb
  • Numbers, not adjectives
    • Having fun doing more, with less
Re: Frequency Divider for older Oscilloscopes??
« Reply #49 on: July 19, 2021, 03:11:08 pm »
I would ask if the waveform was repetitive.

Non-repetitive. You need to capture it alll as it occurs.
There are lies, damned lies, statistics - and ADC/DAC specs.
Glider pilot's aphorism: "there is no substitute for span". Retort: "There is a substitute: skill+imagination. But you can buy span".
Having fun doing more, with less
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf