digital versus analogue oscilloscope

[Thunderchild]

I've just bought a digital oscilloscope on ebay. I'm not hugely impressed but then I've been used to the smoothness of analogue and realize that to have a perfect digital scope would require very good hardware costing alot. my first worry is that even with the input grounded there is a lot of disturbance on the screen, and analogue scope would have a straight line, is this normal of digital scopes or is there an issue with it picking up internal noise ?

anyone able to give a breakdown of what one can expect from a digital versus an analogue scope ? I'd like to leave fair ebay feedback rather than jumping the gun plus get to know my machine proporly and non judge it based on my experience with an analogue one.

 

[blueroomelectronics]

Which scope did you buy?

 

[BrownOut]

I don't remember getting anything like that on a digital scope. How are you "gounding" the input?

 

[Thunderchild]

its a Rigol DS1052E

2 channel 50 MHz (although I thinks its a good 20 MHz with digital interpolation and aproximation to get 50 MHz).

I joined the input to the earth crodile clip so there HAS to be 0.000000000000.... V between the scopes input and earth. I think its maybe picking up interferance from power unit ir its fan ?

 

[BrownOut]

its a Rigol DS1052E
I think its maybe picking up interferance from power unit ir its fan ?

You're probably right. What you've created is a loop antenna.

 

[Thunderchild]

You're probably right. What you've created is a loop antenna.

ah but it does this even with no probe(s) attached and the actual unit is built inside a metal case which is then covered in plastic case, I am contempleting putting a screwdriver in the fan for a second to see if it has any effect, problem is I can't open the damn thing or i'll break the warranty label else I'd look into it and mybe see if something can be done

 

[BrownOut]

I would look for other sources of interterrence before I did that. Try simply moving the scope to another location and repeating your earlier experiment.

 

[Thunderchild]

I think I have solved the mystery, its definitely not the 50 Hz main being picked up because if i put the probe near the power cable with a highpass digital filter of 100 Hz enabled I get a pure sine waveform, however after some fidling with the digital filter settings I have found that the slight disturbance is at aroud a frequency of 118-120 KHz as it dissaperes if i enable a highpass filter with a 120 KHz cut off, I would assume that this is the frequency of the units SMPS, I suppose its not too much to owrry about but odly is is present on all settings at the same physical amplitude on the screen no matter what V/div i set, with the measurements enabled ranges from 240 uV at 2 mV/div to 400 mV at 10 V/div, the mutational nature of the disturbing signal is ahem, disturbing although as i get to know more about it I can more easily "mentally2 remove it when making measurements

 

[bountyhunter]

anyone able to give a breakdown of what one can expect from a digital versus an analogue scope ? I'd like to leave fair ebay feedback rather than jumping the gun plus get to know my machine proporly and non judge it based on my experience with an analogue one.

Digital's main advantage is being able to capture and store files and images. But, you have to pay a lot of money to get a dig scope with enough usable bandwidth to see switching transients and pulse edges. And a lot of times, your trace is a bunch of dots.

The other advantage of dig is cheap to build. Front panel has few switches, just buttons endlessly scrolling through menus. On an analog scope, you can grab the switch you want without searching.

As obvious, I use an analog Tek 7904A scope (best scope ever made).

 

[Thunderchild]

well this has plenty of controls, is remarkably easy to master, the main difference i think is a single V/div knob and vertical position knob as you select which channel your working on, having two different colour traces means that both can be viewed at once on top of each other without confusion.

I did have need for storage function and playback and a slow s/div is very handy in some of my projects, for examply i tried to view the outoput of my friends vintage cars regulator on the analog scope but apart from seing that it went up and down a bit i couldn't get a proper waveform as it meant running the trace so slow it was just a dot going across the screen bobing up and down.

is 50 MHz fast enough for switching transients ?

on the whole I'm quite happy but for this blasted interfearance from the SMPS

 

[bountyhunter]

is 50 MHz fast enough for switching transients ?

Depends. Dig scopes specify samplig rate which is NOT the same as analog bandwidth, but it sounds a lot more impressive. The "good" dig scopes we had were around 2 GIG sample rate as I recall, but they still showed dots on fast transients..

A 50 MHz bandwidth is about minimum for a usable analog scope IMHO. The Tek that I use is about 400 MHz analog bandwidth which is about the highest ever made.

 

[BrownOut]

Random repetive sampling can be used on digital scopes to find all kinds of glitches and anomolies on periodic signals. I used digital scopes for years to track down glitches in production IC testers.

 

[Thunderchild]

your Tek can get up to 500 MHz, (looked it up but depends on installed modules) the digital sampling rate of mine is 1 GSa/s can I assume that means 1 billion samples a second ?

In any case I am thinking of leaving negative feedback on this scope as clearly it lacks in design and construction technique, for the interferance to be increasing as the V/div increases (meaning the image of it stays the same size on the screen no matter how much i increase v/div) means that the SMPS signal is getting into the unit at an intermidiate stage after the voltage/div selection circuitry. This has been cleverly masked by applying appropriate filtering for all of the photos onthe listing and in the manual the only indication of it is in images showing calibration where filtering would peobably upset the calibration. yes reference was made to using filters to get rid of uneccesary noise but it was not stated that this comes from within the scope itself and is 1/5 of a div no matter what, it was implied that it could be part of the measured signal

 

[Thunderchild]

your Tek can get up to 500 MHz, (looked it up but depends on installed modules) the digital sampling rate of mine is 1 GSa/s can I assume that means 1 billion samples a second ?

In any case I am thinking of leaving negative feedback on this scope as clearly it lacks in design and construction technique, for the interferance to be increasing as the V/div increases (meaning the image of it stays the same size on the screen no matter how much i increase v/div) means that the SMPS signal is getting into the unit at an intermidiate stage after the voltage/div selection circuitry. This has been cleverly masked by applying appropriate filtering for all of the photos onthe listing and in the manual the only indication of it is in images showing calibration where filtering would peobably upset the calibration. yes reference was made to using filters to get rid of uneccesary noise but it was not stated that this comes from within the scope itself and is 1/5 of a div no matter what, it was implied that it could be part of the measured signal

 

[bountyhunter]

your Tek can get up to 500 MHz, (looked it up but depends on installed modules) the digital sampling rate of mine is 1 GSa/s can I assume that means 1 billion samples a second ?

Maybe so, I was going from memory and I thought the vertical bandwidth on my 7904A was 400MHz. It might be 500M. My point was that sampling rate and bandwidth are VERY different. A dig scope that samples at 1GIG is significantly inferior to a 500MHz analog scope in being able to track a fast rising signal with sharp peaks.

I actually had two scopes at work (back when I had a job) a TEK digital storage scope with 2GIG sample rate and my analog 7904A. I was a dinosaur and everybody laughed at the huge 7904A, but it was significantly better at seeing fast transisents at turn off edge on fast switch converters. The waveforms were true in analog and sometimes not in digital (clipped off peaks). The digital "lunchbucket" scopes are useful for taking scope shots onto a floppy, but honestly I don't assume I know what a waveform looks like until I see it on an analog scope.

The other upisde of owning a 7904A is that you have to be a REAL man to even lift it, which means the thieves who grab scopes off benches just keep right on walking when they see it (we had a chronic equipment shortage and people were always grabbing equipment wherever they saw it).

 

[bountyhunter]

its a Rigol DS1052E

2 channel 50 MHz (although I thinks its a good 20 MHz with digital interpolation and aproximation to get 50 MHz).

No offense intended, I did a search and found one new for $525. That means it's junk. That's about what a really good set of scope probes costs.

Low cost scopes are OK for general purpose work, but they do have problems. The cheap dump I worked at back in 1982 had bought a bunch of the low end "Leader" scopes and they were junk. As you moved the horizontal trace line up and down it curved into a smile at the upper end of the dispaly and did a frown at the bottom. Cheap scopes are a joke, really only good for go/no-go testing, cant be used for taking data.

 

[bountyhunter]

Random repetive sampling can be used on digital scopes to find all kinds of glitches and anomolies on periodic signals. I used digital scopes for years to track down glitches in production IC testers.

They do have some useful features like that.

 

[crutschow]

I have a TEK TDS 210 60MHz digital scope that I've had for a few years. It has a 1GS/s sample rate so it displays fast transients within its bandwidth without dots. It has little discernible noise on its trances with no input.

If the noise you are seeing is 1/5 of a division in height then it's either a poor design or a faulty oscilloscope.

 

[Thunderchild]

If the noise you are seeing is 1/5 of a division in height then it's either a poor design or a faulty oscilloscope.

I think a design issue, the noise is there the same size in the screen at ALL sesitivities so its onviously bypassing the range selector, if you make voltage measurements the voltage of it increases as you increase the V/div.

I've left appropriate feedback for it !

 

[Dean Huster]

Lots of different things to comment upon here.

I use an analog Tek 7904A scope (best scope ever made).

First of all, I like the front panel differences and features the 7904 has over the newer (and short production run of) 7904A. Both 7904 and 7904A are 500MHz using a 7A19 vertical preamp (that's the "9" part of the 7904's model number). The 7800-series (e.g., 7834 or 7844) have a 400MHz bandwidth. Best scope depends upon the application, though. If you need storage, then the 7834 is the best scope; if you need true dual-beam operation, the 7844 is the best scope; if you need more than 500MHz bandwidth, the 7104 is better; if you need some digital capability, the 7854 is better; if you need super-portability, you have to drop down to a 475, 212, T935 or 335; if you need battery operation, it takes an Option 7 portable and 1106 battery pack or one of the 200- or 300 Sony/Tek-series. and on and on.

You can NEVER equate bandwidth and sampling rate. You can have a 10GS/s sampling rate and a 1MHz bandwidth; or a 500KS/s sampling rate and a 2GHz bandwidth. Either specification would be a horrid and/or wasteful combination -- not unlike using a Peterbuilt road tractor to pull your pop-up tent camper or a VW Rabbit to pull a 12-bottom plow.

Digital devices, by their very nature, including DSOs, have an inherent ±1 digit bobble which means that even the quietest signal you measure is going to "look funny" if you're used to the straight-line trace of an analog scope. Analog folks find that objectionable. On the other hand, high-bandwidth scopes such as Tek's 465, 475, 485 or the 7000-series have an internal scan expansion mesh inside the CRT to help with geometry and trace brightness with the down-side of making the trace fuzzy as compared to their older scopes (e.g., 453, 454 or the 500-series) which had fine, sharp traces. We had customers who sent in their scopes for warranty repair for a "thick trace" (it was still well-within specifications) and we'd swap out the CRT just to make them happy (and save the offending CRT as a new replacement part).

This is why an 8-bit DSO is a pretty crappy "instrument". Only 128 points on the screen allows for some pretty "thick" bobble. 9 bits would be 256 points; 10 bits would be 512 points; 11 bits is 1024 points. By the time you're at 10 and 11 bits, the bobble is about as noticeable as trace width -- i.e., you don't really see it.

I have a TEK TDS 210 60MHz digital scope that I've had for a few years. It has a 1GS/s sample rate

I would have sworn the TDS 210 (60MHz) and TDS 220 (100MHz) had 2GS/s sample rates.

everybody laughed at the huge 7904A,

The other upisde of owning a 7904A is that you have to be a REAL man to even lift it,

I thought the 7904A was light. Compare with a 7704 (not 7704A). It had a non-switching supply and was easily twice the weight of a 7904. the 7000-series overall lightened up the Tektronix product line over the older 500-series. If you wanted a workout for the gonads, you tried lugging a 556 around the shop. It was portable because it had handles.

Dean