Frequency down converter to extend scope's range

[petert]

Hi,

Many SDRs use a frequency up-converter to extend the range to lower frequencies.

Are there down-converters, such that you could look at higher frequencies (maybe up to 500 MHz), though your scope is limited to 100 MHz?

[joeqsmith]

I'm not sure out side of buying a faster scope.   I made something like this many years ago, using a flash converter to capture the waveform then playing the wave back at a slower rate.  Basically, this is designing the whole scope less the display.   Now days, making something like this vs just buying a faster scope makes little sense.

[james_s]

It's not possible to convert the frequency while maintaining the original shape of the signal. Since the purpose of a scope is to see the shape of the signal that wouldn't do you much good.

[DaJMasta]

It's not possible to convert the frequency while maintaining the original shape of the signal. Since the purpose of a scope is to see the shape of the signal that wouldn't do you much good.

This.  A mixer will downconvert the content just fine, but it's not possible to get the whole bandwidth of a 500MHz input into a 100MHz scope that way - you'd just get the 100MHz worth you select of that 500MHz based on your chosen LO frequency, so while you could use a 100MHz scope to capture an 80MHz channel's data waveform with a downconverting mixer (and this is done), you're throwing away all the frequency content outside of the bandwidth of the scope, so the resulting waveform won't look like it would when captured at full bandwidth.


The reason radio applications can do this is because the information content of the signal is centered around a specific frequency, so when you downconvert from this frequency, you move the entire signal of interest into the lower band and with some filtering you can directly digitize it.  Since a scope is a broadband instrument interested in all of the frequency content of the signal, it wouldn't really be valid for normal scope operation and use cases to just downconvert as it wouldn't represent the signal present at the 'input'.

[ogden]

Depending on how good (FFT) spectrum display mode of scope is, one can make it poor man's spectrum analyzer. At least for narrowband ISM applications.

[tonyalbus]

i dont trhink you can do that for a scope, it is usualy done for frequentie counters and its calles a preselector if i am correct.
for a scope you want to keep the shape, for a frequency counter its just a pult to count, what even shape.

Tony

[ogden]

i dont trhink you can do that for a scope, it is usualy done for frequentie counters and its calles a preselector if i am correct.

It's prescaler. Other name - divider. Do not apply here in this discussion because completely different thing than frequency downconverter. Preselector is bandpass filter right after antenna - if we talk about RF receivers.

Yes indeed you can do frequency downconverion for scope - yet you shall know what you are looking at. For example if we take 866 MHz signal with 1KHz deviation and downcovert it into 100KHz "IF" frequency, it will still have 1KHz deviation. Using FFT function of the scope we essantially will look at spectrum of "downconverted" TX signal. That could be useful for those who do not have spectrum analyzer on hand. So idea is not that bad actually. Poor man's spectrum analyzer so to say.

[petert]

The reason radio applications can do this is because the information content of the signal is centered around a specific frequency, so when you downconvert from this frequency, you move the entire signal of interest into the lower band and with some filtering you can directly digitize it.  Since a scope is a broadband instrument interested in all of the frequency content of the signal, it wouldn't really be valid for normal scope operation and use cases to just downconvert as it wouldn't represent the signal present at the 'input'.

I know that I would still have a limited bandwidth of 100MHz just like an SDR. It would still be helpful when analyzing radio signals and looking at generated waveforms of transmitters.

Is there a downconverter like this?

[ogden]

It would still be helpful when analyzing radio signals and looking at generated waveforms of transmitters.

It is already said that after downconversion you do not see generated waveform. Downconverted signal has same modulation properties like phase and/or frequency variation, but it is not original waveform anymore!

Is there a downconverter like this?

Yes. Inside many superheterodyne radios. Usually they are narrowband.  In case you want to build one, check ali/*bay for "35-4400MHz USB Simple Spectrum Analyzer RF Signal Source". It has LO and mixer inside, thou (again) low freq and narrowband IF. It is not that hard to tap mixer output, introduce own IF filter and amplifier. Oh, and you need preselector filter as well.

[edit] Search internet for HAM bandscope or panadapter mod. Radio amateurs are using their receivers as downconverters to look at receive spectrum at IF tap. They use SDR for that, but you may use FFT of your scope as well.

[vk6zgo]

It can be, and had been done.

Not all signals have a bandwidth extending from zero frequency to whatever the maximum frequency of interest is.
Many, although quite wideband, are centred around a much higher frequency, which cannot be seen by the
oscilloscope.

Wideband down conversion can allow you to look at a segment of the specrum, but in the time domain.
Test receivers for analog TV can demodulate say, a 500MHz signal so you can look at tne baseband video signals, or have an IF output, where you can observe the modulation envelope at around 30-odd MHz.

Even though these are not "general purpose" 'scope uses, they are legitimate ones.

[Andreas]

Hi,

Many SDRs use a frequency up-converter to extend the range to lower frequencies.

Are there down-converters, such that you could look at higher frequencies (maybe up to 500 MHz), though your scope is limited to 100 MHz?

That is exactly what sampling oscilloscopes are doing:
DIY example for a sound card here:
https://www.analog.com/en/analog-dialogue/articles/turning-pc-sound-card-into-sampling-oscilloscope.html

with best regards

Andreas

[ogden]

That is exactly what sampling oscilloscopes are doing:
DIY example for a sound card here:
https://www.analog.com/en/analog-dialogue/articles/turning-pc-sound-card-into-sampling-oscilloscope.html

Incorrect. This is equivalent time sampling, not downconversion. Some scopes indeed have downconverters inside. One that I know: Teledyne-LeCroy LabMaster 10-100zi 100GHz Oscilloscope. Further info: The Signal Path youtube video #45.

[joeqsmith]

I never know anymore if my posts are censored or if I actually forgot to press enter.   

If I didn't explain it well enough,  the flash converter would digitize the data and record it to a large buffer.   Once I had the data, I would then play it back at a much slower rate so the scope could keep up.  In my case, I was using analog scopes so I had to refresh the display by playing the recorded data over and over.  The recorder provides a trigger output to the scope to keep it in sync.   There was a delay adjustment for the trigger that allowed me to scroll through the data.     It was very crude but at the time, high speed digital scopes were fairly expensive. 

The setup was very crude but it did allow me to view signals that were much faster than the scope could normally handle.   I built a much slower, higher resolution system years later.  This system had a power amplifier, a few preprogrammed waveforms and a PC interface.   My first arb was born...   

[petert]

It is already said that after downconversion you do not see generated waveform. Downconverted signal has same modulation properties like phase and/or frequency variation, but it is not original waveform anymore!

I am pretty sure we mean the same. When you adjust the scales on your scope and you assume your signal has a limited bandwidth, you'll get the same reading (=waveform) on the screen (within some errors bounds, obviously), even if your original signal is a higher frequency (but, again, limited in bandwidth).

Yes. Inside many superheterodyne radios. Usually they are narrowband.  In case you want to build one,

I was aware that radios had such signal processing stages built-in, but I hoped there would be some commercial options available, similar to upconverters for SDRs.
Ideally, you could put them between the BNC inputs of a scope, and the signal source.

[petert]

Wideband down conversion can allow you to look at a segment of the specrum, but in the time domain.
Test receivers for analog TV can demodulate say, a 500MHz signal so you can look at tne baseband video signals, or have an IF output, where you can observe the modulation envelope at around 30-odd MHz.

Even though these are not "general purpose" 'scope uses, they are legitimate ones.

Thanks, that's exactly my goal. Are you aware of any commercial options that do that (ideally with a selectable band)?

[ogden]

When you adjust the scales on your scope and you assume your signal has a limited bandwidth, you'll get the same reading (=waveform) on the screen (within some errors bounds, obviously), even if your original signal is a higher frequency (but, again, limited in bandwidth).

Same example: 866MHz, 1KHz deviation and 1K symbols/sec is downconverted to 10MHz (IF) frequency. Tell me what good will be for you to look at 10MHz sine waveform coming out of downconverter? Sine will be sine no matter - you look at 10MHz+1000Hz or 10MHz-1000Hz waveform. Hint: RF engineers mostly use spectrum/signal analyzer rather than scope.

[nctnico]

When you adjust the scales on your scope and you assume your signal has a limited bandwidth, you'll get the same reading (=waveform) on the screen (within some errors bounds, obviously), even if your original signal is a higher frequency (but, again, limited in bandwidth).

Same example: 866MHz, 1KHz deviation and 1K symbols/sec is downconverted to 10MHz (IF) frequency. Tell me what good will be for you to look at 10MHz sine waveform coming out of downconverter? Sine will be sine no matter - you look at 10MHz+1000Hz or 10MHz-1000Hz waveform. Hint: RF engineers mostly use spectrum/signal analyzer rather than scope.

It could be interesting to look at how the transmitter is starting up on, maybe look at some phase/frequency shifts in the signal. In theory you should be able to capture the waveform and demodulate it using software on a PC. All in all there are plenty of interesting time-domain properties of an RF signal.

[ogden]

In theory you should be able to capture the waveform and demodulate it using software on a PC.

Nah. Get RTL-SDR dongle and consider it done. For those who work with wider than 1MHz bandwidth signals scope with downconverter won't do any good anyway.

All in all there are plenty of interesting time-domain properties of an RF signal.

Sure. Please share your experience - tell how do you do RF work using scope.

[vk6zgo]

In theory you should be able to capture the waveform and demodulate it using software on a PC.

Nah. Get RTL-SDR dongle and consider it done. For those who work with wider than 1MHz bandwidth signals scope with downconverter won't do any good anyway.

Why do you say that?
There are very few 1MHz bandwidth oscilloscopes around these days.
There is no reason why a "chunk" of spectrum equal to the bandwidth of the associated 'scope cannot be downconverted.

Such tricks were not uncommon in the days of quite low bandwidth oscilloscopes.
Just such a device was described many years ago in an ARRL Handbook.

The idea was so you could look at AM, SSB, & Morse code to check modulation percentage with the former, & the shape of the keyed signal with the latter.

This still could be useful with, say,  a UHF transmitter & a low-ish bandwidth 'scope.

With the improvement in bandwidth of 'scopes, & the advent of, first, analog sampling oscilloscopes with bandwidths into the GHz region, followed by normal DSOs & GHz versions of them, the requirement (which was never mainstream) went away.

All in all there are plenty of interesting time-domain properties of an RF signal.

Sure. Please share your experience - tell how do you do RF work using scope.

Certainly, on one occasion, an oscilloscope found a problem in an FM Broadcast transmitter which appeared to be quite mystifying.
The FET Driver PA drove the grid of a Tetrode output tube, with a meter showing the output of the former.
The normal tuning method was to tune the tube grid circuit for a maximum on this meter, till in passing, we discovered that the Transmitter output increased when we detuned from that "grid peak", instead of falling.

I looked at a monitor point associated with the meter & tuned as normal, displaying 3 or 4 cycles of the 100MHZ signal on a IWATSU 200MHz Oscilloscope.

The problem was immediately obvious, from the shape of the waveform.
When "peaked" the driver amplifier produced a large amount of 2nd harmonic distortion, which gave the waveform a "peakier" shape, with the narrower peaks higher in amplitude.
The meter, being a "peak & hold" circuit, happily interpreted that as a higher output.

The PA stage tuned circuits rejected the harmonics, so the output power dropped.

A Spectrum Analyser would have shown the problem up, too, but the 'scope was there to hand, with a minimum of setup required, & the time domain result of a waveform change was probably easier to associate with the high meter reading than a bunch of extra carriers.

This obviously wouldn't have been an application for the original idea we were discussion, but it is a real example of the use of an oscilloscope for RF work.

[ogden]

In theory you should be able to capture the waveform and demodulate it using software on a PC.

Nah. Get RTL-SDR dongle and consider it done. For those who work with wider than 1MHz bandwidth signals scope with downconverter won't do any good anyway.

Why do you say that?
There are very few 1MHz bandwidth oscilloscopes around these days.

1MHz is RTL-SDR limitation, but scope won't help much because wideband signals usually are high speed transmissions that uses M-ary modulation,  looking at such *waveform* using scope is pointless because in given bandwidth it looks like noise. Using scope as RF digitizer for spectrum/signal analysis PC software can be very good idea, yet I don't know any implementation. AFAIK there is no scopes in the GNU Radio hardware list (yet). Yes, there are Tek MSO5000 and alike which indeed is scope for RF work - with proper RF frontend and spectrum & vector signal analysis software (option).

[Wuerstchenhund]

In theory you should be able to capture the waveform and demodulate it using software on a PC.

Nah. Get RTL-SDR dongle and consider it done. For those who work with wider than 1MHz bandwidth signals scope with downconverter won't do any good anyway.

Why do you say that?
There are very few 1MHz bandwidth oscilloscopes around these days.

1MHz is RTL-SDR limitation, but scope won't help much because wideband signals usually are high speed transmissions that uses M-ary modulation,  looking at such *waveform* using scope is pointless because in given bandwidth it looks like noise. Using scope as RF digitizer for spectrum/signal analysis PC software can be very good idea, yet I don't know any implementation. AFAIK there is no scopes in the GNU Radio hardware list (yet).

We do use scopes for a lot of RF work, for example for looking at (admittedly non-standard) communication signals, simply because a scope gives us a lot more analysis BW than any SA (high end real-time SAs go up to 1GHz, while with modern scopes we get >50Ghz of BW), and while the signal (to some extend) does resemble noise there are certain signal properties which are of interest.

I actually see a lot more use of scopes in RF work in general, for tasks that in the past were performed with spectrum analyzers. SAs are still needed when RF performance is crucial but many modern high-end scopes perform really well in a wide range of RF related tasks.

It's also very likely also the reason why the new Keysight MXR comes with optional RTSA functionality.

Yes, there are Tek MSO5000 and alike which indeed is scope for RF work - with proper RF frontend and spectrum & vector signal analysis software (option).

The Tektronix MSO5000 is the MSO model of the old DPO Series and it has nothing which would make it particularly well suited for RF work (in fact, it's an 8bit scope which is pretty slow, and even more so with SignalVu running on it). You might be thinking about the old MD3000 and MDO4000 Series which *did* have a separate RF input and built-in "spectrum analyzer" functionality but in reality this was just using scope FFT, with limits in analysis BW and resolution as well as a woeful RF performance, and in case of the MDO3000 with a complete lack of phase information and no way to run scope mode and "spectrum analyzer" in parallel.

[ogden]

Mentioning hi-end 50GHz bandwidth scope in discussion where we talk wbout downconverter to reach 500MHz is blasphemy Regarding Tek "RF scopes" - most likely there are better MSO's than one I found using search. Please suggest then, preferably reaching said 500MHz, as low cost as possible.

[edit] 50GHz frequencies are mm-wave, not even microwave or for god's sake RF. You most likely work on radar or UWB application where completely another set of tools, skills and methods are needed. Nothing even close to classic superheterodyne RF stuff (upconversion/downconversion) we were talking here.

[Wuerstchenhund]

Mentioning hi-end 50GHz bandwidth scope in discussion where we talk wbout downconverter to reach 500MHz is blasphemy

You are absolutely right. I shall self-flagelate immediately 

Regarding Tek "RF scopes" - most likely there are better MSO's than one I found using search. Please suggest then, preferably reaching said 500MHz, as low cost as possible.

There aren't a lot of scopes in the 500MHz class which are particularly well suited for RF work (and those that are are also expensive). Most of the scopes are more basic, embedded-type scopes which have no decoding capability, so you mostly have to rely on external processing.

Some scopes, like the MSO5000 (which isn't exactly cheap, though), are supported by signal analysis software of their manufacturers. For example the Keysight DSOX3000A/T, which is supported by the 89600B/C VSA software. However, while the scopes themselves already aren't exactly cheap, the software is pretty expensive, too.

Of course there's always the 2nd hand market. For example, an Agilent DSO8064A or DSO8104A (although the 8064 can easily be converted into a 8104 by removing a resistor so you get the same 1GHz BW) or one of the older XP based 54800 Series scopes, together with a copy of the Agilent 89601A VSA software, would make a good option (provided you can 'find' the VSA software). Getting a DSO8k for cheap (i.e. below $2k) takes a bit of luck but the 54800s can sometimes be found below $1k. The VSA software can run on the scope so a separate PC isn't necessary.

There also are the LeCroy LC, WaveRunner2 LT and WavePro 900 Series, which have 500MHz models and which offer good FFT and some demodulation/signal analysis capabilities which can be 'hacked' (unlocked). LC and WR2LT often go for notably below $1k. Keep in mind that these are primarily analysis scopes not high update rate scopes, though.

There's of course always the option to couple any decent scope (even the ones from Rigol and Siglent) with external 3rd party software. There's SciLab/Xcos and Octave, both open source and free alternatives to Mathlab, which can be used to either use one of the available analysis toolboxes or write your own. These aren't the only tools, there's Mathlab (there's a hobbyist license but it's somewhat pricey), LabView Community Edition, and so on. It requires an external PC and monitor, though.

At the end of the day it comes to what you want to achieve, what your budget is and if you can write some code yourself.

[vk6zgo]

Mentioning hi-end 50GHz bandwidth scope in discussion where we talk wbout downconverter to reach 500MHz is blasphemy Regarding Tek "RF scopes" - most likely there are better MSO's than one I found using search. Please suggest then, preferably reaching said 500MHz, as low cost as possible.

[edit] 50GHz frequencies are mm-wave, not even microwave or for god's sake RF. You most likely work on radar or UWB application where completely another set of tools, skills and methods are needed. Nothing even close to classic superheterodyne RF stuff (upconversion/downconversion) we were talking here.

Funny thing, nobody told the hams who use  superheterodyne principles on 47 & 122 GHz!
https://www.wia.org.au/newsevents/news/2019/20190318-1/index.php

[0culus]

As a historical note, the old HP 8405A vector voltmeter can be used for exactly this purpose. On the rear panel is a pair of BNCs that give you a 20 KHz sampled output that preserves the important characteristics of the RF waveforms (particularly their phase relationship, relative to channel A which is the phase reference). The reasoning behind this is that ultra wideband oscilloscopes were rare and and exceptionally expensive even for many labs back in the 1960s (at that time, the Tek 519 might have been the only "commodity" option IIRC). You might instead have had access to a much lower bandwidth scope, but it could easily work at 20 KHz yet still be able to look at RF waveforms up to 1 GHz when used with the 8405A.