Is Bandwidth/memory depth a waste of money in oscilloscopes?

[Fungus]

Excellent points David. One of the functions I change quite often is to look at the DC/AC component of a test point ie changing the input coupling. I found it a bit clumsy when changing that in each channel vs what I'm previously used to which is just flicking a switch.

Isn't that what 4-channel 'scopes are for, so that you can display both things simultaneously?

[David Hess]

Excellent points David. One of the functions I change quite often is to look at the DC/AC component of a test point ie changing the input coupling. I found it a bit clumsy when changing that in each channel vs what I'm previously used to which is just flicking a switch.

The trigger slope control is also commonly missing.

By the same token I'm probably being a bit critical. Technology wise it's a little like winning the lottery, the complaining when the winnings are delivered in $10 bills instead of $100s I am still amazed at what that little machine is capable of doing, it's just that I'm glad I didn't throw the old Tek out when I had the chance!

I get annoyed when the new oscilloscope cannot, and I really mean cannot, make the same measurements as the old one or performs more poorly.

Perhaps the 2232 performs additional functions but, according to its manual, the waveform storage looks no different than a modern DSO saving the actual data (either in .csv or a proprietary format) to a pendrive and being able to perform advanced analysis on it, or keep it in memory to serve as a reference waveform for simple visual comparison or automated pass/fail tests.

It was mostly intended for storing and displaying reference waveforms and it works very well for this.  The 2232, and 2230 with the option board, even backup the waveform reference memory.  Unfortunately they cannot display the reference waveforms in analog mode.

Brick-wall filtering? If you know how to make a nice brick-wall filter in hardware then I'm sure Tek., et., al. will be very interested.

They are not brick wall but for a while Tektronix high end oscilloscopes implemented switchable 4th order Gaussian filters in hardware.  These days they support something similar in software to backup the low order analog filters.

Excellent points David. One of the functions I change quite often is to look at the DC/AC component of a test point ie changing the input coupling. I found it a bit clumsy when changing that in each channel vs what I'm previously used to which is just flicking a switch.

Isn't that what 4-channel 'scopes are for, so that you can display both things simultaneously?

So the designers intended to replace easy access to the input coupling settings with more channels?

[Fungus]

Isn't that what 4-channel 'scopes are for, so that you can display both things simultaneously?

So the designers intended to replace easy access to the input coupling settings with more channels?

I don't think that was the explicit intention, no, but it's a nice workaround for people not in the "Two channels is enough for me!" camp.

[GeorgeOfTheJungle]

Flipping a switch is easier though, than having to reach for the other probe and swap the GND crocodiles. Ohh, and the trigger settings too.

[Fungus]

Brick-wall filtering? If you know how to make a nice brick-wall filter in hardware then I'm sure Tek., et., al. will be very interested.

They are not brick wall but for a while Tektronix high end oscilloscopes implemented switchable 4th order Gaussian filters in hardware.  These days they support something similar in software to backup the low order analog filters.

Fair enough, but the basic front end still needs higher bandwidth to make these filters work.

Getting back to the point: The assertion that bandwidth isn't related to rise time is false.

With some 'scopes, you need to you know the filter type as well as the raw analog input bandwidth but you can still calculate the expected rise time. All that changes is the constant in the equation.

The assertion made was that this is impossible, that rise times are completely unrelated to bandwidth.

[Fungus]

Flipping a switch is easier though, than having to reach for the other probe and swap the GND crocodiles.

Yep.

Opening a menu and using the stupid scroll knob to select between DC/AC/GND is bad design in anybody's book.

For basic functions like coupling it should be a single button press, even if that means shoving all the other menu items down to make space for an extra entry (one button for AC/DC, another one for GND on/off).

[nctnico]

I get annoyed when the new oscilloscope cannot, and I really mean cannot, make the same measurements as the old one or performs more poorly.

When it comes to DSOs the perfect one simply doesn't exist. I have several DSOs and each has it's own specific area of excellence (not taking bandwidth into account) and price doesn't seem to be the main differentiator here.

[Gyro]

Flipping a switch is easier though, than having to reach for the other probe and swap the GND crocodiles.

Yep.

Opening a menu and using the stupid scroll knob to select between DC/AC/GND is bad design in anybody's book.

For basic functions like coupling it should be a single button press, even if that means shoving all the other menu items down to make space for an extra entry (one button for AC/DC, another one for GND on/off).

Ironically, the Owon USB scopes have a much more minimalist and intuitive UI.

Ironic because:
1) They're  cheap
2) They have far more screen area available to waste than their resolution restricted, small screened cousins.

Maybe they should start shipping with a stylus and / or a USB keyboards and some of the Agilent ones used to do. A simple resistive touch screen and plastic stylus would hardly break the bank. 




   
    (1) Clicking on the small arrow directly brings up the measurement cursors menu for time and / or voltage.

    (2) Clicking on the small '+' icon (only appears when the mouse pointer is in the measurements area) immediately brings up the measurements selection menu.

    (3) Clicking on either of the channel numbers immediately brings up the channel settings menu.

    (4) Volts/div selection, Clicking brings up a normal popup+slider However if you just hover the mouse over it instead you can use the scroll wheel to increment/decrement directly.

    (5) Clicking here toggles between DC / AC / GND for that channel.

    Additional, hovering in either of the channel boxes also brings up a small 'x' icon to turn off the channel. Clicking in a greyed out channel box will turn it on again.

    (6) Clicking in the 'T' box brings up a slider for trigger position. It's normally easier just to drag the red trigger position flag at the top of the screen, but the 'T' box click also brings up a useful 'Reset' button to restore to the middle.

    (7) Time/div... Similar operation to the V/div. Clicking brings up the pop-up/slider but mouse-over and scroll wheel changes the value directly.

    (8 ) Clicking toggles the trigger source between Chan1, Chan2 and Ext (trigger marker on right hand edge of screen changes color to match).

    (9) Clicking toggles trigger polarity, works on edge, slope and pulse (icon changes to match trigger type). In Video trigger mode it toggles Odd, Even, Line etc.

    (10) Click to alter trigger level - Actually it's far easier and more accurate to drag the trigger marker on the right hand side if the screen, but clicking here also brings up two useful buttons to 'Reset' the trigger level and 'Set to 50%'.

- <Space>  brings up (or closes) the last menu selected (like clicking the Menu button)

- When in numeric popups (V/div, T/div, number key will cycle through values for that number. eg. Pressing 2 will cycle 2us 20us, 200us, 2ms etc.

- It will take the first letter of a menu item to select, eg. Trigger mode will take E, S, V, P etc.

(Cribbed from my post in another thread).

[David Hess]

Maybe they should start shipping with a stylus and / or a USB keyboards and some of the Agilent ones used to do. A simple resistive touch screen and plastic stylus would hardly break the bank. 

I am not convinced that a touch screen is a solution but a desk space wasting keyboard (and mouse) are definitely not in an integrated instrument.

HP went for the "single master rotary encoder" interface on their instruments at one time and Tektronix almost followed them down that rat hole in the 1990s.  Did nobody learn anything about interface design from that?

[GeorgeOfTheJungle]

Speaking of desk space, it's true that size matters, in this case smaller is better and modern DSOs win by a wide margin. Good old CRT scopes take up to perhaps 3x as much desk space.

@Fungus, the Rigol 1000z UI is specially annoying because it shares all the channels' control knobs.

[Pete F]

The reality is that a 'scope with 3% error is just a scope with 30% error (ie. gaussian filter) that's been derated, has a software filter applied internally, and a different label stuck on the front.

Incidentally I didn't "massively underrate" the bandwidth that is printed on the front, in fact I specifically stated "The oscilloscope can (and probably will) continue to display signals way above the quoted bandwidth."

Seems like you're not sure which side you're one.

nb. Even a Rigol DS1054Z will display something up to about 300mHz. It'll be very very attenuated but it will be there.

BTW my background was largely in RF, and I'm offering this emoji to the assertion that filtering needs to be done in software 

That's not what I said at all.

Gaussian filtering obviously doesn't need software - we spend a lot of time and money trying to eliminate it in our hardware.

Brick-wall filtering? If you know how to make a nice brick-wall filter in hardware then I'm sure Tek., et., al. will be very interested.

It seems you are becoming increasingly desperate to "win" an argument and are now resorting to pretending that people have said something, then argued against it. Bizarre! For the record it was actually YOU who said that "a 'scope with 3% error is just a scope with 30% error (ie. gaussian filter) that's been derated". I have maintained all along that an oscilloscope will measure well beyond its bandwidth and am well aware of conventional filtering.

It seems very clear that you have never been a professional in this field. I only mention all this as it appears the forum has quite a few people new to the field, and the concept of bandwidth, rise time, and various other oscilloscope functions is something that seems a little confusing for somebody buying their first oscilloscope. I would strongly suggest they be guided by manufacturer's documents in this regard, and I feel the documents I linked to explain the concepts very well. There are many other documents also on the internet, mainly from EEE sources.

As far as using 4 probes, all set to something different and then swapping them around as required, that would have to be the most stupid thing I've heard in a long time. What Dr Google seems to have forgotten is the trigger will be on the wrong channel when you swap probes. So the assertion is that you'll put one probe down, pick up another, change the trigger, do the measurement, move to the next point, pick up the third probe, change the trigger again, etc etc etc. If you're going to change the trigger source, then why not just change the coupling and use the one probe  Too funny!

When working through a board or even a single IC fault finding you'll be swapping input parameters around all the time, and being able to flick between basic functions is essential for any oscilloscope. For those buying an oscilloscope and doing this type of work I think it would be an important thing to check that you're happy with the interface before buying (or just use an old analogue 'scope ). For those doing more development work where the parameters are more fixed, a multi-input oscilloscope would probably be more suitable. I made that point pages ago, as I came from the former camp where speed is king.

I agree that I'm not a huge fan of touch screens, but have become more used to them over the years. They simply replace buttons in any case, and whether the multi-function button is on the screen or a button next to it really doesn't make much difference in my opinion. Anything that is multi-function will require more key strokes to obtain that function when compared to a dedicated switch, but that's just the nature of the beast I'm afraid.

[CatalinaWOW]

User interfaces are a matter of personal preference.  Look at all the flavors of Linux, all of which their designers passionately feel are better than others, sometimes with a rational basis but most often not.

Listen to the opinions of others for insight into what might float your boat, but this is one area where test driving is really important.  It is disheartening to have spent serious money on an instrument that turns out to drive you batty.

[Pete F]

Brick-wall filtering? If you know how to make a nice brick-wall filter in hardware then I'm sure Tek., et., al. will be very interested.

They are not brick wall but for a while Tektronix high end oscilloscopes implemented switchable 4th order Gaussian filters in hardware.  These days they support something similar in software to backup the low order analog filters.

Fair enough, but the basic front end still needs higher bandwidth to make these filters work.

Getting back to the point: The assertion that bandwidth isn't related to rise time is false.

With some 'scopes, you need to you know the filter type as well as the raw analog input bandwidth but you can still calculate the expected rise time. All that changes is the constant in the equation.

The assertion made was that this is impossible, that rise times are completely unrelated to bandwidth.

 

The assertion made, and I would know this is it was ME who made it, was: "The rise time of an oscilloscope is also normally completely independent of the bandwidth, which is why the rise time is also quoted in the oscilloscope specifications."

I highlighted the part you seem to be so confused by. You have substituted independent for unrelated, and got yourself confused with the whole concept. They are independent in that the rise time of various oscilloscopes with the same bandwidth may vary, and the reason why that is so was provided by various documents from the manufacturers. In the past that was possibly much less likely to be so, however with the advent of "relatively" affordable modern very high bandwidth times have changed, and it's something that in some instances potential buyers may need to look more closely at.

With all due respect, there's a time to know when to stop digging and accept that you've misread or misunderstood something. That's your personal business of course, but I'd respectfully suggest just accepting that and move on

[David Hess]

Speaking of desk space, it's true that size matters, and in this sense modern DSOs win by a wide margin, only that in this case smaller is better. Good old CRT scopes take much more desk space, perhaps 3x as much.

But I can stack my CRT oscilloscopes and with their bail, the space under them is available as well.

@Fungus, the Rigol 1000z UI is specially annoying because it shares all the channels' control knobs.

I do not mind that so much as long as it is one button to select each channel and a full set of channel controls are included.  The first microprocessor controlled oscilloscopes shared one set of horizontal controls for multiple timebases so this is hardly new.

What I do not accept is sharing between the vertical, horizontal, and other controls.

[GeorgeOfTheJungle]

What pisses me off is that if you hit, say, the channel 4 button twice it turns off, so when trying to select the channel you want the knobs to act on, it's easy to turn it off instead by mistake, With independent controls that can't happen.

[Fungus]

As far as using 4 probes, all set to something different and then swapping them around as required, that would have to be the most stupid thing I've heard in a long time. What Dr Google seems to have forgotten is the trigger will be on the wrong channel when you swap probes. So the assertion is that you'll put one probe down, pick up another, change the trigger, do the measurement, move to the next point, pick up the third probe, change the trigger again, etc etc etc. If you're going to change the trigger source, then why not just change the coupling and use the one probe  Too funny!

Now you're the one inventing things.

You said you enjoyed the ease of switching between AC/DC on analog 'scopes.

I simply said you can look at both on a DSO with 4 channels. I never said anything about "put one probe down, pick up another, move to the next point, pick up the third probe". Just connect two probes, see AC and DC simultaneously. Trigger doesn't change.

[Fungus]

The assertion made, and I would know this is it was ME who made it, was: "The rise time of an oscilloscope is also normally completely independent of the bandwidth, which is why the rise time is also quoted in the oscilloscope specifications."

Where I come from the word "independent" means there's no relationship, that they can have any value at all.

That's not true, they're normally closely correlated.

The Tek document you linked mentions 'scopes with rise times of 0.45/bandwidth. The Keysight document you linked mentions 0.5/bandwidth. 0.5/bandwidth is only 42% better than the cheapo $350 Rigol DS1054Z.

Prove me wrong! If I'm persistant it's only because I want to be wrong. I love being wrong, that's why I try so hard to be so.

What's the biggest deviation from 0.35/bandwidth you know of in a production oscilloscope? Is there one with even 50% deviation?

Even 50% deviation wouldn't make it "completely independent". "Completely independent" is when I can choose any number I want.

Just admit that and move on.

[Pete F]

OMG, are you serious? Mate you're wrong, you've been shown to be wrong, you admit to be wrong by quoting a 50% error, maybe that's "close enough" for you, but that doesn't mean that will be suitable for others. If you "like to be wrong" I guess today is your lucky day! 

I'm done with this topic. I was asked to substantiate my assertion with evidence to support it, and I have done so numerous times from the manufacturers themselves. If you want to continue to invent things, claim that it's all just a conspiracy from the manufacturers, or that people should use probes without a clue as to how they even work, then knock yourself out. I would however suggest that if you used some of that time to actually do some of this stuff your opinion may be swayed somewhat.

From a more useful discussion perspective, I very much agree with David, it seems that usability has somewhat been compromised in trying to squeeze as many functions in to the units as they can. Unfortunately I think that's a bit of a sign of the times though. With so much equipment being software based it's relatively easy to add "functions" by merely adding code, but that doesn't mean that there has been any huge thought being put in to actual use. Many people will purchase sight unseen (as I did), and if brand X oscilloscope offers more "features" than brand Y, there's a good chance all other things being equal the brand X will outsell brand Y.

Edit: Oh and BTW "Fungus", people like you will always try to have the last word. Every forum has people like that and this one is clearly no different. Fill you boots with the predictable BS comeback to have the final say, I have better things to do with my time than argue idiotic semantics.

[KaneTW]

A const./f relationship is still a relationship. Just because the constant factor changes doesn't mean the values are independent. You can't calculate the rise time from the bandwidth, but there's still a general trend. No one has been arguing that rise time always is 0.35/f or whatever.

[Pete F]

A const./f relationship is still a relationship. Just because the constant factor changes doesn't mean the values are independent. You can't calculate the rise time from the bandwidth, but there's still a general trend. No one has been arguing that rise time always is 0.35/f or whatever.

The constant factor changes and (with modern equipment) is generally unknown these days, therefore they are independent. That is a basic rule of maths.

[Fungus]

The constant factor changes and (with modern equipment) is generally unknown these days, therefore they are independent. That is a basic rule of maths.

At best, the rise time is what a mathematician would call a "limited dependent variable".

OMG, are you serious? Mate you're wrong, you've been shown to be wrong, you admit to be wrong by quoting a 50% error, maybe that's "close enough" for you, but that doesn't mean that will be suitable for others.

(the best example you've shown so far is only 42% more than 0.35  but I'll let that pass)

What that statement means is that if you tell me the bandwidth of an oscilloscope, I can tell you the rise time to within 25%, without any other information.

To me that indicates that rise time is closely related to bandwidth - it only deviates from it by a maximum of 25%.

I'm done with this topic.

"Done with", apart from the post you made after that one. 

[KaneTW]

No, an independent variable would be completely unrelated to the other. That's not the case.

Rise time is Theta(1/f), where Theta is the Landau-theta --- https://en.wikipedia.org/wiki/Big_O_notation

[Mr. Scram]

Yep.

Opening a menu and using the stupid scroll knob to select between DC/AC/GND is bad design in anybody's book.

For basic functions like coupling it should be a single button press, even if that means shoving all the other menu items down to make space for an extra entry (one button for AC/DC, another one for GND on/off).

The DS1054Z isn't known for its refined user interface. I'm not sure that's what you can expect from such a cheap yet capable device either. Analogue oscilloscopes costed one or two orders or magnitude more in their day, so they'd better be well thought out. If course, interfaces are a confusing mix of preference and what you're used to too.

[Fungus]

No, an independent variable would be completely unrelated to the other. That's not the case.

An example of "independent" would be the sample memory:

If you tell me the bandwidth of the oscilloscope it tells me nothing at all about how much memory it has.

That's independent.

[Fungus]

The DS1054Z isn't known for its refined user interface. I'm not sure that's what you can expect from such a cheap yet capable device either.

It doesn't cost extra money to manufacture with a different menu layout. Stuff like the example above means they simply don't care.