to buy or not to buy an Oscilloscope?

[paulcav]

I'm on a budget of AU$400 to purchase an Oscilloscope. My requirements is to develop electronic computer related devices and perhaps fault find computers in general. My dilemma is do i get a DSO (Digital Storage Oscilloscope), as dave would say "for those playing at home",  or a logic analyser.

I have seen a Rigol DS1052E for AU$429 which is a good price for me.

Here is a site thats selling the DS1052E

http://www.rigoloscilloscope.com.au/wholesale/Rigol-Oscilloscope-50MHz-DS1052E-1G-SG-1M-3-years-warranty-Special-Offer-24.html

edit changed to logic analyzer

[TimFox]

A spectrum analyzer (which displays the input signal as a function of frequency) is usually irrelevant to computer-related devices and computer fault-finding.  Did you mean "logic analyzer"?

[tggzzz]

I designed and built my first computer with LEDs, switches, and an analogue multimeter. It worked.

Yes, understanding, imagination, thought, and careful step-by-step construction and testing were required - but that's no bad thing.

But have a look at the Digilent Analogue Discovery. It has limitations, but with the student discount it is difficult to beat.

[guenthert]

I'm not convinced that a (low end) oscilloscope is of much use in developing computer related hardware.  It probably also depends on what kind of computer hardware is meant.  Arduino (and similar) MCU stuff?  Then you probably will want at one time or another a device which can decode SPI and I2C serial buses.  Retro computing stuff from the dawn of the micro computer, then you might get away with a logic probe, but a used (cheap now) multichannel logic analyzer might save quiet some time.  Contemporary x86/ARM stuff, then it will be high end LA and oscilloscopes which are beyond the means of most hobbyists.

I'd wait until you run into a concrete problem which you can't debug with the tools you have or easily can make.

Said that, what you find with an oscilloscope might be illuminating in unexpected ways and perhaps you get a taste of the analogue world ...

[paulcav]

thanks  Timfox i did mean to say "logic analyzer" so i corrected the first post to say logic analyzer

[cheeseit]

The Digilent Analog Discovery is worth considering. If you find that you need a scope, consider saving up or finding the money and go for the DS1054Z instead.

[nctnico]

I don't see how a DS1052E is any better than a DS1054Z for a small price difference. However you might find a DS1052E second hand for a low price. Lot's got sold when the word got out they where hackable.

[james_s]

Given the clock speeds in modern computers, most oscilloscopes are not going to be much use there. A scope can be indispensable but if you're not sure whether you need one, you probably don't, at least not yet. It's important to decide exactly what you want to do and what signals you want to look at. If you're looking at stuff like SPI and I2C you might be well served by a small PC based logic analyzer or mixed signal scope. For lower budget even something like the Bus Pirate can be a useful tool.

[boffin]

The Digilent Analog Discovery is worth considering. If you find that you need a scope, consider saving up or finding the money and go for the DS1054Z instead.

I'll second this, it's a great little tool, plus there's an even lower cost (less functionality) version on kickstarter right now.  Unless you're doing 50MHz signals, this will likely do most of what you need.

There's some more info about it in this thread: https://www.eevblog.com/forum/crowd-funded-projects/kickstarter-digilent-'openscope'/

[tggzzz]

Given the clock speeds in modern computers, most oscilloscopes are not going to be much use there.

The clock speed is irrelevant; the important feature is the transition time. The rule of thumb is that the scope+probe's bandwidth BW > 0.35 /tr, where tr is the rise or fall time.

So, depending on "the computer", 10MHz is unlikely to be useful for signal integrity and glitch detection.

A scope can be indispensable but if you're not sure whether you need one, you probably don't, at least not yet. It's important to decide exactly what you want to do and what signals you want to look at. If you're looking at stuff like SPI and I2C you might be well served by a small PC based logic analyzer or mixed signal scope. For lower budget even something like the Bus Pirate can be a useful tool.

All very true, and I wish more people thought in those terms

[eugenenine]

My requirements is to develop electronic computer related devices and perhaps fault find computers in general.

Be more specific on 'computer related devices'.  Now a days that means something that plugs in via USB as most of us are not designing PCI cards.   
Finding fault in modern computers means swapping pci cards or memory ,etc with a known good one.  Unless you really want to get down to board level repair were you need more than just a scope, you need the tools to rework the components and someone willing to pay fore than what the board is worth in labor costs.

If your finding faults in say a retro computer then most any scope will do.

[Fungus]

Given the clock speeds in modern computers, most oscilloscopes are not going to be much use there.

The clock speed is irrelevant;

No it isn't. A clock speed measured in gigahertz needs a rise time measured in picoseconds. A 100MHz 'scope is useless for that.

the important feature is the transition time.

Transition times are (loosely) related to clock speeds.

[tggzzz]

Given the clock speeds in modern computers, most oscilloscopes are not going to be much use there.

The clock speed is irrelevant;

No it isn't. A clock speed measured in gigahertz needs a rise time measured in picoseconds. A 100MHz 'scope is useless for that.

Sigh.

The transition time determines the highest frequency in the signal, and hence the required scope bandwidth. Full stop.

(If your transition time is too long for two transitions to be completed within a clock period, then you simply don't have a clock!)

the important feature is the transition time.

Transition times are (loosely) related to clock speeds.

No, they are not. There are limits, but that is all. The crude limit is that f_ck < ¹/_2tr, and that's all.

[james_s]

This is getting nitpicky. Fact is for getting a useful view of signals in a modern PC you are going to need a very high bandwidth scope. You're not going to see anything but noise if you try for example looking at the DRAM data bus or PCIx bus with a 100MHz CRO.

[nctnico]

The transition time determines the highest frequency in the signal, and hence the required scope bandwidth. Full stop.

Not really. It depends on what aspect of a signal you are looking at. If you are looking at a 100MHz SPI bus with protocol decoding a 200MHz scope is more than enough because it has enough bandwidth the get the fundamental clock frequency. Ditto for the timing relationship with the other signals in the SPI bus.

[tggzzz]

This is getting nitpicky. Fact is for getting a useful view of signals in a modern PC you are going to need a very high bandwidth scope. You're not going to see anything but noise if you try for example looking at the DRAM data bus or PCIx bus with a 100MHz CRO.

Of course, because the transition time is too short. You'll also need to have good probing technique and appropriate probes!

It isn't "nitpicky" because too many beginners think that "I have a digital signal with a period of 100ms => 10Hz, => so slow I don't have to think about signal integrity".

[paulcav]

this is getting off topic take your transition time debate to another thread such as "to transition time inter-dimensional wormhole argument" for more 

[tggzzz]

this is getting off topic take your transition time debate to another thread such as "to transition time inter-dimensional wormhole argument" for more 

This is a public forum and you don't own the thread. Thread drift has been occurring since the early 80s; deal with it.

[JPortici]

 

25. Relation to the original topic decreases with every single post.

Hm. In my experience you can diagnose and fix problems with far less demanding tools than those required for designing (this assumes that the device you are inspecting worked in the first place)
Have you ever looked at louis rossmann's videos? he doesn't have this astounding equipment, but he can get around it by using his brain a lot.

To quote jim williams

A 50MHz oscilloscope cannot track a 5ns rise time pulse, but it can measure a 2ns delay between  two  such  events.

I hope you start seeing the point

[MrW0lf]

According to my testing 100MHz scope did keep up with very sharp square much further than sine (-3dB @ ~200MHz):


And quite insane timing resolution also (<2ps statistical error):


So with repetitive signals / timing you can do quite a lot with some effort, until analog part passes at least something thru and triggering keeps up. Of course its all purely experimental and needs more proof in practice.

[jacklee]

Good testing, are you an algorithm engineer? 

[MrW0lf]

Good testing, are you an algorithm engineer? 

Thanks. By education I'm software developer. EE stuff is just hobby fun & games. Think I'm accuracy nut... Most satisfying stuff ever watched is machining aluminum on my CNC  But cannot fully exclude just missed parties good enough to compete...

So what can say... if you want to miss out on some hot chicks having fun in a pool - get an oscilloscope