Oscilloscope recommendations

[sirket]

It's quite surprising which 'scopes have the proper display algorithm and which don't. The Instek ones do, as do Tekway. Rigol and cheap Agilent stuff (which is made by Rigol) don't, but the big surprise was Tektronix which also don't display properly! Tektronix have a well deserved reputation for quality instruments, but it seems like they have really dropped the ball on this one!

All of the Tektronix scopes with a MB of memory or more have WaveInspector which allows you to see a proper representation of the entire waveform- including transient spikes- at the top of the screen. WaveInspector makes it trivially easy to find what you're looking for, and even faster to get to it. I've never found a situation in which it failed to show me even the smallest transient.

My guess is that they are playing catch-up with the Chinese manufacturers who have the advantage of starting out with FPGA designs that are suited to this kind of processing where as Tektronix use less generalised hardware that makes processing of large memories difficult. The arrival of affordable large memory 'scopes has forced them to offer similar features but they had to bolt them on to the existing hardware.

The MSO/DPO 2000 series was introduced in 2008 and was designed with full access to a FPGA's if they chose to use them. The particular problem you're describing has nothing to do with FPGA's though- the Tektronix scopes are capable of doing lots of calculations on their stored waveforms. If they wanted to utilize a different sampling mechanism they could easily do so. That said- I've spent a lot of time using WaveInspector and no changes are required- it does a better job that anything else I've used- including, recently, the Rigol scopes.

In those cases where you want more control, memory, or processing power- then you use Labview/Signalview on your nice big computer screen. All of the Tek scopes I'm aware of work with this software, and they do so effortlessly

[iv803]

hi everyone, looks like a lot of you had experience working w/ an oscilloscope and did buying one from a dealer before. I just recently developed a test equipment community to share experiences. You can research and review products / dealers there. I hope you guys wouldn't mind to share your product and dealer reviews w/ our users as well. This will help others researching for test equipment products and dealers. Thanks!

URL: http://www.testunlimited.com/

[AdHoc]

I'll admit I rarely stray into the analogue world unless pushed I recently purchased a USB based scope from http://www.cleverscope.com/. So far I've liked the results, the integration with a handful of digital probes is quite useful. I usually use an Intronix USB LA (http://www.digitalgraphics.com.au/intronix/index.html) for the digital but I have found that whenever I do need to check an analogue source it's almost always been related in someway to a digital line and this tie in with the scope helps. For the cost you can prob. get a decent stand-alone scope but I decided to try and reduce my bench clutter and get a USB version that looked like it covered most bases. There's also a fairly limited second hand market in Australia so that helped with the decision too.

[Entropy]

Cleverscope - $1100 for only 100 MSa/s???? That's absolutely pitiful.

Rigol DS1052E is fully standalone, $400 currently, 50 MHz analog bandwidth and 1-2 GSa/s.

[fat16lib]

The Cleverscope is an interesting choice. With 4M sample storage and 10-14 bit digitizers it is totally different than the other scopes discussed here.

A typical scope like the Tek TDS2022B has 8-bit digitizers and 2.5K sample records.

Could you give a bit more detail how you are using it?

[mojo]

14 bit ADCs don't appear to be that useful in most situations, unless I am missing something. Most (all?) DSOs can scale the input via amplifiers or switching DACs so you can measure very large and very small signals with the same 8 bit resolution.

The only time you would need 14 bit is if you had a signal where you needed to analyse a signal on both a large and small scale simultaneously with one probe and the sample memory. If you could use two probes you could just connect them both to the same signal and set one for the large and one for the small signal. If you could look at separate cycles on each scale you could simply capture one and then the other.

I can't really think of anything off the top of my head that would need it.

[fat16lib]

Digitizing signals to 14 or more bits is often done in labs where I worked. Frequently this is done to preform spectral analysis. This can be a quick test to see if an amplifier distorts a signal.

I can think of dozens of reasons for precision digitization of signals. I spent years working with such systems.

I was just curious to see how Cleverscope was being used. You don't see use of these features in hobby electronics very often. I wasn't suggesting Cleverscope's features are not useful.

[AdHoc]

Yes it is not that hard to find a stand-alone scope with better performance figures than pretty much all USB PC hosted ones. However, as I said it was also an attempt to reduce the bench clutter which to be honest is a bigger problem than the 'scope specs for me at this point (I have a nasty habit of having several projects spread out on the go at the same time).

From a quick check of the USB scopes available, most (other than cheap knock-offs) were priced at around $AU1000 or a little lower. So I was looking for features that set some apart from others, the digital lines are a useful addition, the 14 bits sampling could be useful but not essential, a basic waveform generator may be useful (again reducing clutter , a reasonably large memory for storage was also a plus and a protocol decoder (for the digital lines) was also handy as in some cases that will save me setting up the LA as well. Getting a cheaper one is certainly an option, but I've seen some pretty awful software with those and I wanted something that would work without fuss, as I don't want to have to spend my time fighting bad software. So for about the same cost you could get a stand-alone with better sampling rates, similar bandwidth but far less features in all other areas. Your choice would largely be have to be based on what you expect to be using it for, as there are lots of variations available in that price range.

[mojo]

Digitizing signals to 14 or more bits is often done in labs where I worked. Frequently this is done to preform spectral analysis. This can be a quick test to see if an amplifier distorts a signal.

Ah yes, a good example.

It's an interesting product, no doubt. What is the software like? Can you dump the raw measurements to a reasonably easily processed file format (e.g. CSV, text)?

[ganzziani]

I made a comparison table of hobbyists oscilloscopes under $500.00.
You can filter and sort the table to find the one you want.

[ianryeng]

Just wanted to put in a suggestion for the Hantek/Tekway scopes.

They have higher than usual resolution for a budget scope and can be firmware 'upgradable' to 200MHz with a little effort

The only downside to some people might be that they are 2 channel but I am using the DSO5062B at home and am very happy with it.
In the office we have a lecroy waveace and wavesurfer and I like my Hantek better than the waveace (even the wavesurfer seems like its just a big screen...).

Just a suggestion for anyone looking to buy a scope and don't want a Rigol

-Ian

[mojo]

Spotted an interesting article about oscilloscope probes today: http://www.dfad.com.au/links/THE%20SECR ... 0OCt09.pdf

(warning PDF)

TLDR; They are complex little devices, not at all trivial. Quite how much difference there is between a cheap Rigol probe and a Tektronics or Instek is a subject about which little has been written. For many hobbyists it probably doesn't matter much, but I can think of situations where it does. Might be interesting to see how much difference some more expensive probes on a cheap 'scope makes.

[joeld]

Cleverscope - $1100 for only 100 MSa/s???? That's absolutely pitiful.

Rigol DS1052E is fully standalone, $400 currently, 50 MHz analog bandwidth and 1-2 GSa/s.

Rigol is a great scope, but I think many that buy it are looking fairly narrowly at the problems they need to solve (I have a DS1102E and a bunch of other scopes).

You need to first ask yourself "what am I going to measure with my scope?". A few common areas are:

1.Analog circuits in the audio range. This includes audio. Your sample rate need only be a few hundred KHz here, or maybe 1-2 MHz, but 10 bits or wider ADC widths are very important, as an 8-bit converter has a max SNR of 50 dB or so. Rigol is so-so here. Cleverscope and others crush the Rigol here, however, with the wider ADC widths.

2.Analog circuits in the few MHz range. This includes switching power supplies, basic video. Generally an 8-bit ADC is fine here, although more is better. Scope sample rate need only be 100 Msps or so. Most all scopes with sample rate over 50 to 100Msps will be fine here. However, if you are looking for noise among a legit signal, then the wider ADC width will again help. Rigol's 8-bit is barely passable for this function.

3.Analog circuits in the 10's of MHz range. This includes RF IF, advanced video, timing on USB2.0 full speed. Sample rate here needs to be 200Msps and higher, and ADC width of 8-bits is usually fine. Rigol is uniquely suited here. This is why I have it: For looking at clocks in the 40 to 200 MHz range.

4.Analog circuits in the 100's of MHz range. This includes RF, very advanced video, ethernet, slower (SDR) memory interfaces. 8-bit ADC is width is almost mandatory, as converter cost climbs quickly if you want wide converters and insane speeds (and there are huge export restrictions on this technology too). This is squarely in the realm of high-end Agilent and similar, thought a few chinese scopes are making headway here.

5.Analog circuits in the GHz range. This includes HD video, high serial interface (GBit interfaces such as used on LCDs), timing on USB2.0 high speed, timing on 800 MHz DDR, etc. This is squarely in the realm of high-end Agilent and similar.

6.Digital signals in the ~5 MHz range and below. These come from all sorts of smaller controllers (PIC, AVR, MSP) running at 20 MHz and below. This is easily captured on simple logic anlyzer such as USBee or Saleae.

7.Digital signals in the 10's of MHz range and below. These come from 32-bit processors running at ~50 MHz, such as ARM M3. This is easily captured on basic USB logic analyzer such as QuantAsylum or Intronix.

8.Digital signals in the >100 MHz range and below. These come faster ARMs such as ARM9. This is easily captured on advanced USB logic analyzer (such as Intronix, but not QauntAsylum).

9.Digital signals in the GHz range and below. These come from the fastest mobile and desktop processors today. Again, very specialized and very expensive tools.

So, think more broadly about your problem set. You can find a lot of scopes out there that can do better at solving a wider range of problems. Oscilloscope are not stereo equipment. You don't just go after watts and total harmonic distortion. Look closely at what you are trying to solve and pick accordingly.

[mojo]

Excellent post joeld.

I have a Saleae Logic which I have found very helpful for microcontroller work, especially because it can capture an effectively infinite number of samples. When debugging some comms that suffer from intermittent issues I can just grab 30 seconds worth of data at 24MHz and pick through it on screen.

I have done quite a bit of video work with my Instek 100MHz 'scope, including debugging timing issues. I am starting to work with ARM cores now running at 80MHz, although the maximum I/O clock on many of them is around 20MHz. One thing to keep in mind is that in many cases you can slow down I/O or buses to help debug them if your equipment is struggling at full speed, particularly if you control the clock as with SPI and many parallel interfaces. The Logic will however be inadequate at 20MHz (rule of thumb is you need to sample at least 2x as fast as the signal you are measuring, preferably 4x), which is possible on a 20MHz AVR. I'd say that if possible get a 100MHz 'scope rather than the only slightly cheaper 60MHz ones because that will get you up over 4x the maximum signal frequency, except in the case of XMEGAs which can hit 32MHz.

Also beware of manufacturer stated sampling rates. For example my Instek is a 100MHz 'scope and in my tests will actually cope reasonably well at 130MHz (the fastest signal I have had a chance to test), but the sampling rate is listed a 1 gigasample/second. When you look closer that is the oversampled rate on a repeating signal, i.e. the display is actually a combination of several cycles. Most of the time that is fine but if you were looking for a specific non-repeating artefact it wouldn't work very well.

We have already covered the problems some 'scopes have with displaying large sample buffers so I won't go back over that old ground.

I have done audio, video and switch mode PSUs in the 500KHz range without any problems. One thing I would suggest if doing power supplies is make sure you have good well insulated probes and always use a multimeter to check voltage levels first. An old analogue 'scope bought cheap from a junk sale is ideal actually, because you won't be worrying about breaking an ancient £30 wreck with only one working channel and some CRT burn in

[joeld]

These are good points you bring up re: safety and slowing down the system and streaming. All viable considerations too.

For safety, if I worked a lot on anything near or above mains, I'd only trust a scope with CAT II written on the front panel. If it doesn't have CAT II on the front, then the manufacturer didn't intend on you to stick the probe into anything more than 40V or so. An experienced user can get by with less, but only in a pinch.

Slowing down the system is very effective as long as you aren't facing signal integrity issues. But it is indeed a very good trick you bring up.

I don't have the Sealea, but I have a USBee SX which is similar and can let me capture infinite streams of digital. That is another dimension that I missed and is a very good point you make: There is no substitute for streaming if you are trying to figure out what went wrong in a long data xfer.

My current arsenal of test test equipment in order of usefulness is:

* QuantAsylum QA100 for most all mixed signal SW/HW debug and protocol interpreting
* Intronix LogicPort for complicated and fast FPGA problems (usually SDRAM interface issues as it has a 500 Msps)
* Rigol DS1102 for quick checks and clock integrity above 25 MHz
* USBee SX for FGPA configuration issues as these are slow but require you to examine a long stream to find the problem.
* Parallax Scope too basic and used as a loaner these days for friends in need
* Link Instruments MSO19 too basic and used as a loaner these days for friends in need

If you look at the list above and ignore the last two, they all do something the others can't do. But the Rigol isnt' at all near the top of the list in terms of what is needed most for me. I think many buy the Rigol not knowing exactly what they need. Instead, I'll bet they'd find a good mixed-signal scope much more useful.