70 MHZ Oscilloscope with higher rated probes

[Rick King]

Hi,

My first post and a retired beginner for hobby.
First, I searched the web and here for a previous post so not to duplicate (i.e. eevblog.com: probes on Google.) So, sorry if I did.

So you know my use: Only Hifi amplifiers, Tape decks etc. Keep in mind I am a newbie and learning, although I retired from IT. I may be way off base in my thinking. Reading a lot of books and watching EEVBlog videos. I may be ahead of myself in my learning curve. At my age I am trying to learn only what I need to know for hobby and probably missing a lot of basics.

I bought a cheap scope (Hantek 70 MHz, although has decent reviews) to learn with. I also purchased extra 200 MHz probes. I'm not sure if I needed them. When looking at SINE waves from a Audio Function Generator the signal seems distorted at x10 and even worse with the 200 MHz probes. [Probably not the best way to get a signal, but using software as opposed to a proper stand-alone Signal Generator, which may be the problem] Lastly, I have calibrated and also tried to fine-tune with the screws on the probe; which helped a little.

I guess my question is guidance. Do I need 200 MHz probes? Should I buy a separate unit? Many of the HiFi techs I talk to say they don't use one that often. Any recommendations for a better learning path would be welcomed as well. Hope I have made some sense.

Cheers!
Rick

P.S. I am not thin skinned. I can handle directness. Just want to learn.

[tggzzz]

Welcome. Thinkg and wanting to learn are a good starting point.

I can't think what you might mean by "distortion"; a photo would help, as might a description of your probing technique.

200MHz probes will be fine,but their extra bandwidth will be "wasted".

You may find it helpful to read the references at https://entertaininghacks.wordpress.com/library-2/scope-probe-reference-material/ particularly the "ABC" guides and the safety guides.

[radiolistener]

wide bandwidth probes are bad for digital oscilloscope with low bandwidth. It's better to use the probe with narrower bandwidth than your oscilloscope input bandwidth. This is because it's better to cut-off high frequency components, because otherwise it will be folded into higher Nyquist zone of ADC and leads to a signal distortion due to aliasing.

For a clean signal which bandwidth completely fits into the first Nyquist zone (half of the sampling rate) you will not notice any problem. But when signal consists of high frequency components you will see significant difference.


Do you measure D-class amplifier? What kind of distortion you're talking about?

[tggzzz]

A probe's frequency rolloff is much too poor and poorly defined to be of any use as an anti-aliasing filter.

The scope should incorporate the specific anti-aliasing filter that the scope needs. If not, then get another scope. (There are rare counter cases, but they are not relevant to a beginner)

[bdunham7]

Describe your Audio Generator and post a screenshot of the distorted waveform.  Also, the exact model of the scope? 

Are you saying the the signal is not distorted using the 1X setting on the Hantek probes?  Have you calibrated the probes using the scopes calibrator output?  What frequency sine wave are you trying?

[Vovk_Z]

Probe bandwidth doesn't matter. It may be 50 MHz or 500 MHz. It only must be calibrated with the exact input. That's all. Wider bandwidth is almost always better then narrow.

[Rick King]

Really appreciate the responses. It must be very frustrating trying to help someone (like me) that has no clue what they are doing, or how to ask a proper or relevant question.

I will respond back fully in an hour with some images.

I will say that I am just experimenting right now (learning). I don't know enough to start testing my stereo components. Connecting directly from an iPad with  EE Toolkit Audio Function Generator. I made new cables to connect directly to the probes rather than several adapters and the signal is cleaner.

Thank you!!

Back in a few...

Rick

[radiolistener]

A probe's frequency rolloff is much too poor and poorly defined to be of any use as an anti-aliasing filter.

yes, probe frequency rolloff is not the best, but it helps to reduce high frequency components to eliminate aliasing issue. Especially for a slow time/div modes which is required to analyze low frequency audio signals. And it's better than no filtering at all when you use very wide bandwidth probe with low bandwidth oscilloscope.

The scope should incorporate the specific anti-aliasing filter that the scope needs. If not, then get another scope. (There are rare counter cases, but they are not relevant to a beginner)

Yes, good oscilloscope should have antialias filter on the input. But using narrow bandwidth probe helps to reject unwanted high frequency components much better.

For example oscilloscope antialiasing filter may reject some component for 30 dB. And narrow bandwidth probe provide additional 30 dB rejection. In total we have 30+30 = 60 dB rejection. When you use wide bandwidth probe you will get just 30 dB rejection, which is not enough to not see unwanted aliasing.

Cheap oscilloscopes, like these 70 MHz may not have antialiasing filter at all.

For example FNIRSI-1013D shows aliasing distortions because it's specification bandwidth is higher than the half of it's ADC sample rate.

But the thing going to be much worse if you needs to measure low frequency signals (this is exactly the case of topic starter) and your oscilloscope has too small memory or too slow DSP (which is guaranteed for a cheap oscilloscopes). In such case oscilloscope needs to force reduce ADC sample rate for slow horizontal resolutions. Because it will be unable to capture and process enough sample count at full ADC speed for a long time interval (which is required to see a low frequency waveforms).

In such case internal antialiasing filter of oscilloscope cannot help, because it's cut-off is much higher than half sample rate for selected mode and your waveform will suffers distortions due to aliasing. This is a standard issue for a slow time/div modes in a cheap oscilloscopes.

Probe bandwidth doesn't matter. It may be 50 MHz or 500 MHz. It only must be calibrated with the exact input. That's all. Wider bandwidth is almost always better then narrow.

this is false information. Using expensive wide bandwidth probe with a cheap narrow bandwidth oscilloscope may leads to a waveform distortion due to aliasing. How much worse depends on the used antialiasing filter and osclilloscope bandwidth to oscilloscope sample rate relation and also oscilloscope memory size and it's DSP/CPU speed. For a cheap oscilloscope it may be very bad and very noticeable.

[Rick King]

I will try to explain what I am thinking in my head so you can help me better. I appologize that I was not clear.

My whole mindset, right or wrong, is setup. Am I using the right probes and setup for learning. I am also reading and watching videos from basic elctronics level. I understand in principle the concept of circuitry and know what all the components are. I am still learning how they work together to form a functional unit (i.e. amplifier). At times I want to get away from the "books" and learn the equipment; and the oscilliscope is what I started playing with lately. Before I spent weeks on a multimeter. I will get back to the multimeter later. I want to know the basics of each of my tools for now instead of staring at them.

I picked up a Hantek DSO5072P 70 MHZ Scope for learning. I have no idea what I untimately need. One of the reviews said to upgrade the probes. I purchased 200 MHZ probes based on the fact that they were more expensive. Don't laugh.

I set the Scope and calibrated probes per the manual. I then wanted to learn about Signal Generarators. Rather than purchase one, I purchased an iPad app (actually paid for it) to learn what it does.

What I found was that with a SINE wave it was fairly clear using either probe. When I switch to SQUARE wave its where most of distortion is. I think my connections were bad when I first posted.

The real question initially was: Does it matter which probes I use; 80 MHZ or the 200MHz. I was thinking I needed the best setup I could get for learning so that I could see what it does. I was concerned that if my basis for learning was off, it would effect knowing what a good signal should look like; for my scope. I have seen what a Siglent looks like.

I could go on, but let me post some images to see my current state.

[Vovk_Z]

About last picture with a squre wave: it is 1 kHz signal, so 70 or 200 MHz BW probes doesn't matter. I guess this signal is so bad at its sourse.

And don't forget to use always 10:1 divider (otherwise any probes will affect a circuit by their capacitance).

[radiolistener]

I then wanted to learn about Signal Generarators. Rather than purchase one, I purchased an iPad app (actually paid for it) to learn what it does.

What I found was that with a SINE wave it was fairly clear using either probe. When I switch to SQUARE wave its where most of distortion is. I think my connections were bad when I first posted.

That's the problem of your signal source. iPhone has something like 48-96 kHz sample rate (like any other sound card) and a low pass antialiasing filter on the output, so it cannot produce nice square wave. You're needs a good signal generator in order to get more nice waveforms

[Rick King]

I then wanted to learn about Signal Generarators. Rather than purchase one, I purchased an iPad app (actually paid for it) to learn what it does.

What I found was that with a SINE wave it was fairly clear using either probe. When I switch to SQUARE wave its where most of distortion is. I think my connections were bad when I first posted.

That's the problem of your signal source. iPhone has something like 48-96 kHz sample rate (like any other sound card) and antialiasing filter on the output, so it cannot produce nice square wave. You're needs a good signal generator in order to get more nice waveforms

OK. That makes since. In the future, when I do start using these tools for "HiFI components"; will the SQUARE wave be used?

[radiolistener]

OK. That makes since. In the future, when I do start using these tools for "HiFI components"; will the SQUARE wave be used?

Smartphone as a signal source for HiFi component testing is a very bad choice. I recommend you to buy a good old (used) analog signal generator. If you're interested in just an audio circuits, not HiFi or audiophile, you can buy some cheap Chinese signal generator like FY6800. They have about 45 dB dynamic range. This is not HiFi, but good enough for a regular audio component testing.

Another way for audio component testing is to buy some professional audio card like EMU with a good dynamic range and use special software for PC. It produce specific signal output and then analyze how it was changed on the input and displaying you frequency response and many other measurement details, like IMD distortions level, etc.

[Rick King]

I really, really appreciate you guys taking the time to answer my questions; and yes it is helpful. I feel like a mouse walking around with giants, but gentle giants.

Further comments are welcomed.

Thank you sincerely,

Rick

[bdunham7]

OK. That makes since. In the future, when I do start using these tools for "HiFI components"; will the SQUARE wave be used?

You've discovered Gibbs and Nyquist all at once!  Presumably the output DAC for the iPad is 48kSa/s (48,000 samples per second)

https://en.wikipedia.org/wiki/Gibbs_phenomenon

https://en.wikipedia.org/wiki/Nyquist%E2%80%93Shannon_sampling_theorem

Your scope is working OK from the looks of it.  The bandwidth of the probe is much less important than it's other qualities and how well it is calibrated, especially at audio frequencies.  As far as square waves, yes they can be used to test audio equipment if you have sufficient understanding, but first you have to understand what a square wave actually represents.  Then you have to understand group delay and phase response so that you can understand what to expect from various audio circuits.  I'd suggest you keep it simple for now, especially if you want to use a software signal generator.

[tggzzz]

A probe's frequency rolloff is much too poor and poorly defined to be of any use as an anti-aliasing filter.

yes, probe frequency rolloff is not the best, but it helps to reduce high frequency components to eliminate aliasing issue. Especially for a slow time/div modes which is required to analyze low frequency audio signals. And it's better than no filtering at all when you use very wide bandwidth probe with low bandwidth oscilloscope.

The scope should incorporate the specific anti-aliasing filter that the scope needs. If not, then get another scope. (There are rare counter cases, but they are not relevant to a beginner)

Yes, good oscilloscope should have antialias filter on the input. But using narrow bandwidth probe helps to reject unwanted high frequency components much better.

For example oscilloscope antialiasing filter may reject some component for 30 dB. And narrow bandwidth probe provide additional 30 dB rejection. In total we have 30+30 = 60 dB rejection. When you use wide bandwidth probe you will get just 30 dB rejection, which is not enough to not see unwanted aliasing.

Cheap oscilloscopes, like these 70 MHz may not have antialiasing filter at all.

For example FNIRSI-1013D shows aliasing distortions because it's specification bandwidth is higher than the half of it's ADC sample rate.

But the thing going to be much worse if you needs to measure low frequency signals (this is exactly the case of topic starter) and your oscilloscope has too small memory or too slow DSP (which is guaranteed for a cheap oscilloscopes). In such case oscilloscope needs to force reduce ADC sample rate for slow horizontal resolutions. Because it will be unable to capture and process enough sample count at full ADC speed for a long time interval (which is required to see a low frequency waveforms).

In such case internal antialiasing filter of oscilloscope cannot help, because it's cut-off is much higher than half sample rate for selected mode and your waveform will suffers distortions due to aliasing. This is a standard issue for a slow time/div modes in a cheap oscilloscopes.

If you choose a particularly crap scope to justify your statements, then you are making a mistake.

If you are considering noise at low sweep speeds, then you should consider using an external low pass filter to limit the noise. It will also be useful to consider the scopes quantisation noise at the relevant signal levels

Out of curiosity,how much experience do you have with designing DSP systems?

[bdunham7]

Using expensive wide bandwidth probe with a cheap narrow bandwidth oscilloscope may leads to a waveform distortion due to aliasing. How much worse depends on the used antialiasing filter and osclilloscope bandwidth to oscilloscope sample rate relation and also oscilloscope memory size and it's DSP/CPU speed. For a cheap oscilloscope it may be very bad and very noticeable.

Maybe in some extreme case, but this scope--like most--has a 20MHz bandwidth filter that ought to be engaged when measuring slow signals HF signal noise is any concern.  Aliasing a 1GSa/s scope, no matter how crappy, at audio frequencies, seems unlikely.  As for lower sampling rates due to memory limitations, the square wave he demonstrates works out to a sampling rate of 12.5MSa/s (40K record length, 3200uS on screen)  for a 6.25MHz Nyquist.  The BW filter won't help if there were HF noise, but the probe bandwidth won't matter either.

[Vovk_Z]

Using expensive wide bandwidth probe with a cheap narrow bandwidth oscilloscope may leads to a waveform distortion due to aliasing. How much worse depends on the used antialiasing filter and osclilloscope bandwidth to oscilloscope sample rate relation and also oscilloscope memory size and it's DSP/CPU speed. For a cheap oscilloscope it may be very bad and very noticeable.

Of cause, you are right too, but I can't agree with you totally, because of several reasons.
Yes, aliasing is a weak side of any digital scope, but to fight with it by an unknown amplitude-frequency characteristic of a probe is not reliable.
What you say, as for me, means that for the same signal a digital scope has to have wider bandwith (up to several times) then analog one. This is a drawback - working with a digital scope you always have to take into consideration a possible aliasing.

 A frequency response of a cheap non-brand probe is usually 'wavy' and unknown. So (as for me again) I would rather have a good high-BW probe than cheap, which works like filter but very unknown filter.

[Vovk_Z]

I mean I would better have one onknown variable (aliasing) than two unknown variables - aliasing and low-BW probe.

[Rick King]

Until I buy a "real" signal generator, would I be better off using my PC and the external DAC? It will output up to 32-bit 384khz.

[Kleinstein]

The PC sound-cards often contain a reconstruction filter. So even if the DAC may be 192 kHz the actual BW may still be limited (in the extremes some 22 kHz).
The sampling rate is not enough to make sure to get higher bandwidth output.
For just audio a normal soundcard is fast enough though.

To get some higher frequency contend, one may consider building a simple square wave souce (e.g. NE555 based). The harmonics extened to beyound 100 kHz and the shape of the transitions can tell a little about the higher frequency range - though a bit tricky to interprete for a beginner.

Ususally a faster probe is OK, especially in the low cost range. Even the cheap probes usually come as 100-150 MHz.

[Rick King]

Between the two, the PC DAC looks good in all waveforms up to about 22.0500KHz if I configured it right.

[radiolistener]

If you choose a particularly crap scope to justify your statements, then you are making a mistake.

As I know, almost all budget oscilloscopes with price up to 500 USD have aliasing issue at slow sweep. They are not crap and have antialiasing filter. But they have just 14-28 megapoints memory which is just 14-28 ms. This is not too much for a low frequency audio signals, because 14 ms is a one period for 71 Hz. Many other oscilloscope have even smaller memory.

Out of curiosity,how much experience do you have with designing DSP systems?

Cannot say exactly, as I remember I made my first DTMF recognition firmware on a PIC16F84 microntroller at about 1995 as a student. Probably this was my first project which utilized signal processing. If we're talking about deep signal processing and ADC/FPGA stuff - about 8 years. Actually I'm software engineer, electronics is my hobby, but before I go into software stuff I also completed education for digital electronics and computer design.

Just a month ago I made my own 14-bit oscilloscope with ADC running at 96 MHz (high speed ADC + FPGA + GMII ethernet + software processing on PC), here is screenshot of 1 MHz a square wave output from PSG9080 with no antialiasing filter and with 30 MHz LPF.

[David Hess]

There is nothing wrong with higher bandwidth probes on a lower bandwidth oscilloscope.

Instead of thinking of the oscilloscope probe's bandwidth specification as its -3dB point, consider it to be the highest usable frequency before aberrations in the probe's response corrupt the signal.  This becomes apparent with a high frequency transient response test on a probe which has a lower bandwidth than the oscilloscope; the oscilloscope will display aberrations caused by the lower bandwidth probe.

[bdunham7]

Until I buy a "real" signal generator, would I be better off using my PC and the external DAC? It will output up to 32-bit 384khz.

Yes, give it a try and see how it looks.  Keep in mind that the distortion you saw on your square wave was not a result of the Nyquist limit, but rather of the output filter required to avoid it.  With a higher sampling rate, the transition frequency might not be any higher, but they can use a simple first or second order filter.