Advice before I jump off the deep end

[TheUltraNoob]

Greetings,

I'm slowly putting together some basic equpment for my electronics classes. Right now I'm looking for a nice analog scope to use for both class and to grow into as my skills advance. Seems that that the Tektronix 2225 is good choice but how will it serve in the future?  I'm also looking at the Tektronix 2465B but is that just stupid overkill?  I don't mind spending a little bit more if it's something I can grow into.

Any advice?

Best regards,
The Ultra Noob

[free_electron]

Why analog ?

[vk6zgo]

Why not?

[Mechatrommer]

i suggest getting digital scope as it has memory storage you can save signal for post processing and wider application such as logic, spectrum and signal analysis etc. analog scope is only alot more usefull for analog geeks who hate computers .

[vk6zgo]

Who hates @#$$$^@@###78855!!!!!! Computers!

Both would be nice,but as a retired OF I can't afford a nice Rigol to sit alongside my Tektronix 7613----bummer!!

[EEVblog]

I'm slowly putting together some basic equpment for my electronics classes. Right now I'm looking for a nice analog scope to use for both class and to grow into as my skills advance. Seems that that the Tektronix 2225 is good choice but how will it serve in the future?  I'm also looking at the Tektronix 2465B but is that just stupid overkill?  I don't mind spending a little bit more if it's something I can grow into.

The 2465B is awesome, but if you are spending a decent amount on it then it can be argued that the money is better spent on a modern digital scope.
Getting any analog scope is great for a beginner, but the only scope that you can "grow into" is a digital storage one I'm afraid.
The 246B is exactly the same as 2225, or any other analog scope, just higher bandwidth.
Neither  will give you the jump into the single-shot storage world.

Dave.

[vk6zgo]

Then again,some of us like to see things as they happen,& not wait for the "instant replay"!

[EEVblog]

Then again,some of us like to see things as they happen,& not wait for the "instant replay"!

Sure. But what happens when you want to see something that only happens once? 

Dave.

[poodyp]

Then again,some of us like to see things as they happen,& not wait for the "instant replay"!

If it's happening 500,000 times a second what does it matter if it's "as they happen" or delayed a few ms? No matter what it's happening more times than you could possibly perceive.

[AndyC_772]

Sure. But what happens when you want to see something that only happens once? 

Dave.

Upgrade? Make it happen again?

I know it's a debate that'll rage on for ages, but I much prefer an analogue scope to a cheap DSO that gives a simple, pixelated display. They just throw away too much information to give a true, accurate picture of the dynamic behaviour of a signal.

I think the danger is that, if someone has only ever used a cheap DSO, they won't necessarily understand what it is that they're missing.

[BravoV]

Buy both, a "cheap" vintage but high speed analog scope + a cheap DSO. 

[EEVblog]

I think the danger is that, if someone has only ever used a cheap DSO, they won't necessarily understand what it is that they're missing.

That is why I encourage beginners to get a cheap (or free) analog scope first.
But when it comes down to it, if I had to only have ONE scope in the lab, and had the choice between a cheap 50/100MHz digital 1052E or a high end 1GHz kick-arse analog scope, I'd chose the cheap digital. It is just so much more versatile.
Missing a bit of dynamic info is nothing compared to not being able to capture and analyse single shot events.

Dave.

[jucole]

When i got into electronics 2 years ago I bought a cheap old analogue scope (Kenwood CS-1577A);  Much of the appeal of analogue is I think because you're seeing very much the pure signal;  I learned a great deal from it, very recommended!    But as I progressed it was no use for digital signals so I bought a cheap old HP1660-AS,  the digital scope has been superb!   but now I'm ready to move on.

If I knew then what I know now I would have saved up and bought a "newer" 2nd hand digital scope, perhaps one that has I2c etc and other tricks,  but then I wouldn't have picked so much up and had as much fun along the way.

[vk6zgo]

I think the danger is that, if someone has only ever used a cheap DSO, they won't necessarily understand what it is that they're missing.

That is why I encourage beginners to get a cheap (or free) analog scope first.
But when it comes down to it, if I had to only have ONE scope in the lab, and had the choice between a cheap 50/100MHz digital 1052E or a high end 1GHz kick-arse analog scope, I'd chose the cheap digital. It is just so much more versatile.
Missing a bit of dynamic info is nothing compared to not being able to capture and analyse single shot events.

Dave.

It depends a lot on what you are looking at.

I tend to use a 'scope to look at everything,DC,analog & digital signals,---but mainly when troubleshooting equipment which has worked.

In most cases if I am messing around with digital stuff,the 'scope is really being used as a huge logic probe,just to see if the signals are there.

I understand,that if you are designing,especially with digital stuff,you do need all the special "bells & whistles"of a DSO.

As I said earlier,I'd like to have both,even if I would still probably use the old analog beast more often.

[AndyC_772]

Allow me to illustrate why I feel an analogue scope is still very much worthy of serious consideration compared to an entry level DSO.

When my own scope first came out, the idea of an 'analogue-like' graded intensity display on a digital scope was still new - not to mention expensive - so it has a button on the front panel which allows it to be switched on and off. Tektronix call it "digital phosphor" mode.

With it off, it behaves much the same as many modern budget scopes; it takes one sample per pixel and displays it. With it on, however, it samples at a much faster rate and displays a trace whose intensity at any given point is related to how often the signal being observed is actually at that level. The trace is much more like what a true analogue scope would display.

Here's a capture from a circuit which I currently have set up on my bench, with DPO mode turned off - ie. the sort of trace than an entry level DSO would show:



...and for comparison, here's the same three signals re-captured with DPO mode turned on:



For the two simple digital signals there's clearly not much to choose between the two, but for the analogue signal on ch 2 there's a world of difference. With the analogue-like DPO mode turned on, you can see the regular short duration spikes that the scope simply misses without it because they're not as wide as a full pixel with this time base.

What you also don't see is the difference between how stable the two traces are. With DPO mode on, the trace is rock solid, and virtually identical on every sweep - but without it, the short spikes in praticular come and go with each sweep, dancing up and down depending on whether the scope happens to be sampling at the instant they're there or not.

In 'simple' digital mode, the trace just doesn't really represent what the analogue signal is doing, and that's why I'd rather have a good used analogue scope than an entry level new DSO.

[Mechatrommer]

With it off, it behaves much the same as many modern budget scopes...but for the analogue signal on ch 2 there's a world of difference. With the analogue-like DPO mode turned on, you can see the regular short duration spikes that the scope simply misses without it because they're not as wide as a full pixel with this time base.

in "entry level" rigol DS1052E, its called "peak detect" mode instead of "normal" mode, not color graded though. if you need more sensitive triggering, use sweep "normal" or "single". sweep "auto" may misses some events. just a few tips

if you need color graded, you can code some PC program for it. take many samples and calculate persistent color, slow but at least you see some nice colors to your liking... sorry i should have said in my earlier post... analog scope is alot usefull for those who hate "programming with computers" but as i venture deeper into high speed analog, the need for high speed scope emerged, and its hell expensive in digital world, even still the 2nd hand 500MHz - 1GHz analog scope i ebayed few hours ago

[EEVblog]

For the two simple digital signals there's clearly not much to choose between the two, but for the analogue signal on ch 2 there's a world of difference. With the analogue-like DPO mode turned on, you can see the regular short duration spikes that the scope simply misses without it because they're not as wide as a full pixel with this time base.

But surely they would have been captured with any deep memory scope?, even a low end one like the 1052E
And even if you have a short memory digital scope, that's what Peak Detect acquisition mode is for.
So I think that in this case, your assertion that those peaks couldn't be picked up by a digital scope is not valid.

Some digital scopes won't display glitchy detail like that deep in the memory unless you are zoomed in sufficiently (I guess so they don't "clutter" the display), so it's standard practice with digital scopes to capture the waveform and them zoom in to see if anything is there in the detail.
Just like on analog scopes it's standard practice to turn the intensity knob up to full briefly to see if there is any dim detail you are missing.
Analog scopes aren't magic either, you can certainly miss detail that even a cheap digital scope with Peak Detect mode can see.

Dave.

[AndyC_772]

Guys, I'm just trying to present some evidence to help people make up their minds what they'd most like to see on their scope displays. IMHO it takes experience to know what setting on a DSO is giving the most accurate display for any given waveform, and I think analogue or intensity-graded displays are easier to interpret correctly.

Here's the same traces captured in peak detect mode:



I won't deny that, in this particular case, this mode does indeed show up a feature of the waveform which is interesting and important. It's the mode I'd use. However, I often find it makes signals appear noisier than they are, especially on a high bandwidth scope.

Just for fun, and to show the importance of selecting the right acquisition mode for a given signal, here's the same in hi-res mode:



So, IMHO of course, we have:

- one setting (standard) which is liable to miss brief events completely, or at least show them inconsistently,
- one setting (peak detect) which captures all manner of nasties, and
- one setting (hi-res) which cleans up noise and leaves a fine, sharp trace - but also has the effect of reducing bandwidth so dramatically that a signal which is clearly visible in the other modes can disappear almost completely.

These are the reasons why I personally prefer analogue, even if that means losing the ability to capture single shot events.

[Wim_L]

These days, digital oscilloscopes have become good enough to mostly replace the analog ones. The cheapest digital ones, though, are nowhere close to matching the performance of a good analog scope. Image quality, excellent intensity grading, and dead time between triggers are the main issues. On the truly bad DSOs, there's also the possibility of being misled by bad interpolation algorithms, but that should be uncommon. Aliasing is an issue you need to watch out for too (not a major problem when you know your instrument well, but it can be confusing when starting to learn, so it may be best to always turn peak detect on unless you have a good reason not to on a low end DSO).

On the other hand, digital scopes have advantages for single-shot capture, and low repetition rate signals at very high sweep speeds (though there is the occasional analog scope that will handle those well --- don't expect to find them cheaply though). Then there's the ability to calculate with signals. The typical analog scope will handle subtracting two signals. A digital one will do anything the designers put in its software, often including FFT and the ability to define your own math functions.

Bandwidth? Perhaps not so essential when just starting out. More is better, sure, but for someone just starting out it will be tricky to probe high frequency signals correctly. Anything above 100 MHz probably isn't going to see much use in beginner experimentation.

I'd say, if you can afford a good digital scope, it's probably all you need nowadays. If you're limited to an ultra low budget DSO, a decent second hand analog may be a useful addition for some measurements.

[T4P]

Also, stay away from second hand DSOs, the 250MS-1GS models aren't just old enough to be cheap second hand or common.
Or probably already broken
The DSOs you see are from vintage DSOs
even the small ones with only 2.5K sample memory, yes teks ...
Or even worse ... 20MS for 100MHz input bandwidth 

[jucole]

Also, stay away from second hand DSOs, the 250MS-1GS models aren't just old enough to be cheap second hand or common.

HP1660AS - 136 Ch. 100 MHz State, 500 MHz Timing, Logic Analyser w/ 2 Ch., 250 MHz, 1 GS/s Oscilloscope.
ex MOD. It cost me £75 from Ebay!  works great! although I'm tired now of manually counting I2c clocks and data lines etc.

[T4P]

Well of course, that's just a monster.

[free_electron]

Or even worse ... 20MS for 100MHz input bandwidth 

So ? Agilent's newest 'monster' is a 16 channel , 60GHz analog bandwidth , 1Gs/s machine.
Works perfect ! you can live perfectly fine with a subsampling system IF , and ONLY IF you measure repetitive signals ( for example eye patterns )

These subsampling machines were built for that purpose.

[baljemmett]

HP1660AS - 136 Ch. 100 MHz State, 500 MHz Timing, Logic Analyser w/ 2 Ch., 250 MHz, 1 GS/s Oscilloscope.
ex MOD. It cost me £75 from Ebay!  works great! although I'm tired now of manually counting I2c clocks and data lines etc.

I've been idly wondering if it'd be possible to write a serial decoder for that class of machine using the SDK HP produced for writing inverse assemblers.  I've only briefly glanced through the example code, but it might be plausible; just haven't got around to trying it yet.  (Of course I presume the 1660s have that facility; my 1650B does, and I'm pretty sure it's from a lesser series!)

[jucole]

I've been idly wondering if it'd be possible to write a serial decoder for that class of machine using the SDK HP produced for writing inverse assemblers.  I've only briefly glanced through the example code, but it might be plausible; just haven't got around to trying it yet.  (Of course I presume the 1660s have that facility; my 1650B does, and I'm pretty sure it's from a lesser series!)

I'm sure you could easily do just that, it would take a bit of time though to achieve something fairly polished and robust.   Instead you might want to consider the $30 Bus Pirate http://dangerousprototypes.com/docs/Bus_Pirate