Cheap PC Oscilloscopes - Any Recommendations?

Missionary Man asks: "I'm an electronics teacher looking for a good (but reasonably cheap) PC based oscilloscope for classroom demonstration purposes. I've done a reasonable amount of research and come up with a few contenders. Ideally I'd like something with a bandwidth of up to 40MHz and 2 channels. Does anyone have any tales to tell regarding the use of any of these scopes (or any others I haven't found or mentioned) and can recommend a suitable device?"

"Here's the list of my findings so far:

• The DS2200C from USB Instruments will do 2 channels at 12 bit resolution, but only to 200KHz.
• The PCS100 from Velleman at QKits runs to 12MHz, but only 1 channel. It has a bigger brother, the PCS500, that has 2 channels and 50MHz bandwidth, but is a lot more expensive.
• Picotech do cheaper ones, like the ADC-40/42, but these only operate in the KHz ranges.
• Link Instruments sell the DSO-2102S that runs to 60MHz with 2 channels, but it's a bit out of my price range.
• Finally, I found the bitscope which seems to be just what I'm looking for, combining a 2 channel scope and an 8 channel logic analyzer for a reasonable price.

I'm hoping to spend US$300-$400. I recognize the software is a fundamental part of the successful operation of these units and any comments regarding the bundled programs would be most helpful too!"

Comments

Here's one to check out -Price page is unreachable

[MikeDawg]

Check this one out. . . I don't know the price because their price page is broken. . . It looks like everything you need in a PC Oscillioscope.

winamp?

[VegetariMan]

I think my media player has an oscilloscope...

Re:winamp?

[seanadams.com]

You know, for classroom demonstrations a sound card should be perfectly acceptable for all kinds of things... especially considering that there is also very nice free spectrum analyzer software available. On top of that, students can easily repeat experiments on their systems at home. Of course with a max sampling rate of 48KHz there's only so much you can do, but low speed osciallators, R/C circuits, switching power supplies and all sorts of things can still be seen to a useful extent.

Re:winamp?

[FredGray]

You know, for classroom demonstrations a sound card should be perfectly acceptable for all kinds of things...

My understanding is that sound card inputs are AC coupled, so you won't be able to see anything much slower than about 20 Hz. That might be OK, but it's not quite an oscilloscope.

Re:winamp?

[seanadams.com]

Good point... I think you could probably address that by bypassing the cap on the input, but then your "zero" level would probably be 1.15V.

Re:winamp?

[Rebar]

Depends on the size of the capacitors on the input side of the sound card - the larger the RC time constant, the lower the frequency signal that can be observed. My old SB compatible card could pick up as low as 3hz signals really easily - I had an animometer hooked directly to it and counted zero-crossings to determine the wind speed. The signal may not be that linear under 20hz, but it's still really easy to see.

I agree with the parent post about a sound card being a nice classroom demonstration scope. For higher mhz, you can pick up a real scope from Ebay for much cheaper than a PC card solution.

Re:winamp?

[tftp]

You described a stroboscopic (sampling) oscilloscope. However it has a serious, theoretical flaw - the bandwidth of the signal is still limited to the 20 kHz that the sound card can take in.

What you are doing is basically undersampling the incoming signal and then assuming that the original falls into one of many aliases that the undersampling generated.

In other words, if the signal changes while you are sampling and reconstructing it, the change is lost and results in incorrect reconstruction.

The original poster's question assumes that the PC-based scope is the best solution to his problem. My EE experience tells me that generally you want a standalone scope (you want as many screens as you can get, space be damned.) Teacher's needs, of course, may benefit from the PC-based scope (multicasting the readings to students' computers, for one.)

Bankruptcy Auctions!

[Supp0rtLinux]

Search google for bankruptcy auctions. I'm sure you can find what you're looking for. I work for a biotech that saved 75% of our initial budget by acquiring medical laboratory supplies from six other biotechs that went bankrupt in recent months. We did such a good job, our VC company gave us more than we asked for to buy the rest of what we needed.

The only thing necessary for Micro$oft to triumph is for a few good programmers to do nothing". North County Computers

Re:Bankruptcy Auctions!

[bigberk]

I work for a biotech that saved 75% of our initial budget by acquiring medical laboratory supplies from six other biotechs that went bankrupt in recent months

Six others? OMG, and that didn't in the least suggest to you that maybe you're in the wrong kind of business?

Re:Bankruptcy Auctions!

[The Clockwork Troll]

Six others? OMG, and that didn't in the least suggest to you that maybe you're in the wrong kind of business?

How many IT companies have gone out of business in recent years? Does that dictate that the upstart firms with financial discipline and valuable products/services should pack up shop and close their doors? Because it's an industry many have failed in?

Sorry to go off-topic but since the question was about oscilloscopes I felt I had to add a logic probe!

Re:Here's one to check out -Price page is unreacha

[msgmonkey]

Price page is here but it looks like it's around $900.

Like Skee Lo, I wish...

[ObviousGuy]

I sometimes think back to my college days and really regret that I didn't focus more on the hardware side of computer engineering, primarily delving into the Computer Science land of algorithms and language design. In the last several years of my employment in the embedded space, I have come into contact with more hardware than you could shake a stick at, and without that grounding in hardware that I should have gotten in school I feel a little overwhelmed when faced with anything deeper than a block diagram.

I also wish that teachers like yourself didn't have to worry about providing materials like this within such a tight budget. It doesn't sound like this is just for this year's class, but something that can be used year after year. $300 for a material that can be used multiple times seems very cheap, especially considering the intrinsic value of the tool. Schools should be at liberty to spend what is necessary to bring the classes up to exceptional levels. Considering how the U.S. lags behind most other modern Western nations in Math and Science, such tight-fisting seems to be one significant factor in this drop off.

Good luck in finding the right tool.

Educational device

[Rosco P. Coltrane]

Two thoughts about your question:

- Do you really need 40MHz for educational purposes? Unless this is a device to be used in a college or higher education class , you can display sub-10KHz signals to teach a class how to use a scope.

- When I was at school, I learned how to use a real scope, with knobs and buttons and a not-so-perfect green screen, and I reckon it was way better to touch these dials and controls and have a direct feel for what they did on the screen than set some virtual thing and grab perfect-looking samples, to understand how things actually worked.

In short, any old regular scope that's well explained by the teacher is probably better than any interface+software setup that "isolate" the student from whatever electrical phenomenon he's trying to expose.

Re:Educational device

[Rosco P. Coltrane]

On the one hand, it's cool when kids get to learn with the best equipment.

I'm afraid this is one of the greatest misconception in the world education. It's definitely not cool when kids get to learn with the best equipment : kids should learn the basics on simple, self-explanatory equipment. Complication and better equipment can come later, when the basics are understood.

That's the same reason why, after universities have taught CS students Java, C++ and Visual Basic before C and assembler, and churned out unfinished computer "engineers" for years, us low-level programmers still get high-paying jobs doing the old non-object-oriented, boring un-cool engineering the right way.

Re:Educational device

[Rolo Tomasi]

Scope bandwidth - you can never have enough. If your budget is limited, you're better off with a good old analog scope. Digital bandwidth is too expensive. Calculate 10 samples per Hz of the signal that you're sampling. So, if you're looking at a 40 MHz sinewave, you need 400 MSPS. And that's just to roughly make out the shape of the waveform, 800 MSPS would be better. Just imagine what a single period of a sinewave looks like when it's divided into only 10 discrete levels. Looks like a staircase.

The other thing to consider is the input amplifier bandwidth (this applies both to analog and digital scopes). This is also known as slew rate. It describes how fast the input amp can follow a signal change. Imagine an ideal square wave with zero rise time. It has infinite bandwidth. What does this mean? If the signal has a faster riset time than your input amp, your edges will be smoothed out. If I look at the output of a 40 MHz TTL oscillator (which outputs a squarewave) with my 60 MHz scope, I see what is almost a sinewave. "60 MHz scope" means the scope can display a 60 MHz sinewave. The sinewave is the waveform with the slowest slew rate. All other 60 MHz waveforms will also look like a sinewave on this scope. If you want to analyze square waves, your scope will only show a halfway accurate depiction of the signal if it has upwards of ten times the bandwidth of the signal.

There are also problems when measuring high bandwidth signals. Above about 80 MHz, you need to use BNC jacks on both sides, properly terminated with 50 Ohms, or the stuff you see on your scope screen will have very little to do with the actual signal. Not many outputs can drive 50 Ohms. You need to build special prototypes of your stuff that are intended for scope measurements. You can't just take a scope to your CPU or stuff like that.

If I had to make a recommendation, get a 60-100 MHz analog scope in good condition. Tek 465 is a good model.

If you want to look at digital stuff, get a logic analyzer. There are some interesting DIY projects on the web.

For higher frequency or RF stuff, a spectrum analyser can't be beat. But good ones cost more than a luxury car. If you're really serious, there are DIY projects on the net for that, too.

Re:Educational device

[I don't want to spen]

Very off-topic, but I remember seeing a demonstration 10-15 years ago of the latest Spectrum Analyser, where the salesman made a big deal of the battery backed RAM saving the settings when the device was switched off. One of the older engineers said "we've got that on the analogue spec analysers, we call it a knob."

Re:Educational device

[cybermace5]

A 465 is good, but a 465B is better. The Tek 465 is about 30 years old now, and the 465B is only 20 years old. I was able to pick up a nice 465B with all manuals and accessories on eBay for $100 recently.

If the poster really wants a digital oscilloscope, head on over to fpga4fun.com. There's some neat little FPGA projects, based on a little FPGA board the guy designed and is now selling for $50. One of the applications is a digital sampling oscilloscope; it actually looks pretty neat. With the FPGA board and ADC board, it's pretty cheap too.

Re:Educational device

[Rolo Tomasi]

No, it's not. You don't buy digital scope to analyse perfect sinewaves. When you're looking at complex waveforms that might include single events and glitches, you need at least a sampling rate ten times your signal frequency. The only way you get your original signal back if it's been sampled at or near the Nyquist frequency is through interpolation. That might be ok for audio applications, but when you need to extract secondary information from the signal, there's just no way. Ten times is the bare minimum, and Tektronix advertise their scope bandwidth as 1/10th the sample rate.

Re:Here's a good one

[Anonymous Coward]

This is not an informative comment - the budget is between $300-$400, and the oscilloscope linked is $21,500 list price - I doubt the used price would be within the budget. Mods, let me remind you that moderation is not a race. Follow the link and pay attention before you assume it's informative. I'll gladly take the karma hit for this if the mods resposnible for this can wake up.

3 times the highest frequency being measured

[Futurepower(R)]

Remember that the oscilloscope bandwidth close to the frequency of the waveform being measured distorts that waveform. (In phase if the frequency being measured is a sine wave.) You need an oscilloscope bandwidth maybe 3 times the highest frequency being measured.

ICs often have very high potential bandwidths, and, when something goes wrong, even an audio IC can have sometimes have parasitic oscillations at extremely high frequencies. If you are working on a circuit, you need to be able to see those parasitic signals.

I don't like this fact, because it is expensive, but 100 MHz seems to be a good oscilloscope bandwidth. I bought a very old Tektronix scope to get the needed bandwidth at a reasonable price.

Re:3 times the highest frequency being measured

[Anonymous Coward]

it's *5* times the signal.

see the tektronix documentation, for ex. 2x is the standard shannon-nyquist theory, but to get proper results, go 5x. tek has a huge document on it, quite informative.

Re:3 times the highest frequency being measured

[Anonymous Coward]

I think it's at least 10 times the signal.

Imagine a sine wave.
If you do 2 samples/period, you measure at 0 and 180 degrees and you get a flat line.
If you do 4 samples/period, you measure at 0, 90, 270 and 360 - resulting in triangular waves.
If you do 6 samples/period, you measure at 0, 60, 120, 180, 240, 300, 360 - the result is more like a sine wave, still the amplitude is about 30% off.
If you do 10 samples/period, the result finally resembles the original (to a certain degree).

Who the fuck modded this Informative?

This is flat out wrong.

The sampling theorem is, strictly:
If a continuous function only contains frequencies within a bandwidth B, it is completely determined by its value at a series of points spaced less than 1/(2B) seconds apart.

This means that with anything greater than 2 samples you can theoretically reconstruct the continuous bandwidth limited signal. So, your first example is somewhat right, and is why we say "less than 1/(2B)" rather than "less than or equal to" This means 2.1 samples/period or even 2.00001 samples/period is enough.

You are making a common misconception that the reconstruction of sampled signals is done by plotting points and connecting the dots with a line. This is not the ideal way. Ideal reconstruction looks something like this:
Sigma(i=0, K, x_i*sinc(pi*(t-t_i)/dt)). Of course this is not a practical reconstruction in real time applications because of the need for "future" data, but with it you can mathematically prove to yourself that one can reconstruct perfectly with any sampling rate greater than twice the maximum frequency.

You should read up on signal processing.

The OP was talking about bandwidths as it applies to practical devices which involves stuff having to do with clock jitter and SNR (oversampling can be used to increase effective SNR).

Re:3 times the highest frequency being measured

[sbaker]

THe frequency rating of something like a scope refers to the highest frequency sinewave it can reconstruct. Nyquists sampling theorem says that you need MORE than twice the frequency. If you have exactly two samples for one cycle of a sinewave, you can't reconstruct it. However, if you have 2.1 samples - you can.

You are concerned about reconstructing the details of the shape of the waveform - but that's not at issue here. We're only promising that if you look at a fourier transform of your signal - then the highest component of that that we'll reconstruct must be less than half of our sampling frequency.

When we are way up there where we have only two-and-a-bit samples, the only shape we can possibly see is a sine wave because anything else would have higher frequency components - and then we are off the hook - we don't have to reproduce those kinds of waves accurately.

So - the issue is whether you can deduce the phase, amplitude and frequency of a sine wave from just two-and-a-bit samples. Nyquist says you can - and he's right.

Stop thinking about how you'd draw graphics of a sinewave using only three numbers - this is signal processing - not graph plotting.

If you want to see a 40MHz square wave, you may well need a 400MHz scope - but for a 40MHz sine wave, a 40MHz scope (which samples at a little over 80Msamples/sec) should be plenty.

Re:Try the sound inputs for a demo

[Rosco P. Coltrane]

For cheap demo purposes, consider using the analog sound ports (line-in). I have had good success doing that for lower speed (44.1K).

That's not going to help you display DC and very low frequency signals. In fact, a soundcard's frequency response is un-linear enough that you really don't want to use one for demo purposes.

Comment removed

[account_deleted]

Comment removed based on user account deletion

Not PC

[linuxwrangler]

OK, I know you asked about PC based scopes but if a plain old scope will do the job consider used.

A friend of mine bought a couple at a ham radio swap meet from a guy who buys surplus lots. IIRC they were dual-trace and something like 20MHz (he ended up getting one for me and for several other interested friends).

They were selling for ~$20 which means you could have a scope for every student in a class of 20 and still stay in your budget.

Must it be PC-based?

[Hayzeus]

You can generally get excellent deals on analog, non-pc based scopes on eBay. I use a Tektronix 2235 100mz dual trace unit that only ran me around USD $200 -- and this was a few years ago.

I also have an Ozi-Fox handheld that has a PC and/or palm-based interface. It only does 20mhz and is single trace, but they are fairly inexpensive (< $90.00 USD) -- you could buy multiple units for classroom use. The display on the unit itself is not great, but works well for quick-and-dirty work.

Good luck -- m

Why PC tethered?

[Jeff DeMaagd]

It seems people on slashdot are a bit overeager to buy PC based devices that can't be used as a standalone device. So far that I've seen, the benefits are mostly just cost, at the expense of portability, usability and quality.

I'd just find a real scope on surplus somewhere.

I just have a 'scope on "loan" from a local EE guy. Just an analog one. Effectively it is mostly a gift, but there are times he wants an analog scope so he wanted the understanding that he can get it back on occasion. For most uses, a digital one does fine.

Did you look all that hard?

[Jerk City Troll]

It is not difficult to find a veritable mount of cheap oscilloscopes on eBay. You say you only need it for demonstration purposes, so why do you need something shiny and new? It can even be argued that the older analog oscilloscopes are better than newer digital ones. As always, resort to eBay if you need something not so good, and fairly cheap. Chances are you can find it there.

Audio Frequency Freebie

[G4from128k]

If a 2 channel audio frequency scope is all you need, then Oscilloscope 2.5 for Windows might do the trick. It uses the inputs on a standard sound card to grab 2 channels at up to 20 kHz. Disclaimer: I've not used it and a bit of Googling may find better alternatives.

Its slow, but is free (assuming you have the sound card).

Windoze luzers with Tektronix 2400-series DSOs....

[John Miles]

... can download free/open-source plotter emulation software at http://www.qsl.net/ke5fx/misc/7470.zip. This may be helpful to you if you buy an older scope from eBay. It will let you grab screenshots, overlay them, print them, and save them in several formats including their original HP/GL-2 plotter language. You can see some typical screen captures at http://www.qsl.net/ke5fx/synth.html.

I use a Tek 2430A on my own bench. These are great scopes -- you can get 150 MHz bandwidth for about $400-$600. A National Instruments GPIB adapter to interface it to the PC will set you back another $100.

I'm trying to add support for as many instruments as I can to this package. Any interested parties should feel free to email me...

Try the USRP

[gustaffo]

If my understand is correct, some of the guys from the GNU Radio Project have developed a USB based software radio device that works with in linux. It is called the Universal Software Radio Peripheral. I think the first prototypes have shipped. The cost is pretty close to your price range. You can see it in action running an oscope program here. And of course it can be extended to do many more exciting things.

Re:Try the USRP

[dfranks]

Except that afaik it will not provide any triggering capability (other than software pseudo-trigger). This may be partially offset by the ability to record really long sample sequences.

Bitscope

[GigsVT]

Since no one who's replied has answered your question, instead choosing to talk about unrelated things, I have to say that I'd go with the Bitscope.

Visit #electronics (our electronics+open source channel) on irc.freenode.net if you want to discuss.

Re:Here's a good one

[John Miles]

Quite a few scopes by both Tektronix and Agilent run Windows these days. Windows can be a reasonable embedded platform. It provides a consistent user interface, saving the equipment manufacturer a lot of software work that's outside their core competency. System stability isn't too bad since the OEM maintains tight control over the drivers and only run one application.

Not a bad solution, all in all, at least until you hook up an Ethernet cable to transfer some plots and your $20,000 scope gets r00tz0red and drafted into service in the war against SCO.com.

nyquist frequency

[jonniesmokes]

Missionary Man asks: "I'm an electronics teacher... ...Ideally I'd like something with a bandwidth of up to 40MHz and 2 channels.

The bitscope only has a 40MS/s data aquisition rate. Assuming that that's for both channels - 20MS/s each, then your left with a nyquist of 10Mhz. And you really need to oversample a waveform a lot more than x2 to see what it looks like. The analog bandwidth of the bitscope is high, but the A/D conversion will result in a lot of aliasing. That said. Its a really impressive unit for $400.00. I didn't think you could find something nearly that fast for under $700. Not exactly what you're looking for.

The software for a scope is pretty important - but without the raw A/D speed and resolution you won't get very far.

Re:Educator?

[NoseBag]

Ya left one out. Write/go to any local company doing e-tech type manufacturing or engineering. Ask if they'll donate their old obsolete eqpt that they've already written off. I've never been in an engineering lab yet that didn't have at least one dinosaur in it.

Re:Here's a good one

[cscx]

Don't make me tell you about the time a lab full of $70k Tek logic analyzers for a class I took had to be shut down because the douchenozzle sysadmins put the LAs on the Internet with external IPs (no firewall even), yet never thought of patching them. They all got infected with one of the worms.

Re:MHz vs. MSPS

[GigsVT]

You don't need 10x samples for most things. You need around 5x to catch most transient events. You only need 2x to make a waveform per Nyquist.

Most digital scopes do a sort of interpolation when the frequency is above Nyquist, they will sample the signal for several periods and reconstruct what the waveform looks like, this works with any repeating signal, but when you use this you can't catch transient events, for that you do need a sampling rate several times the base frequency.

Try DRMO for some HPIB/GPIB compatible gear

[pvera]

DRMO = Defense Reutilization Marketing Office

This is the military agency that sells surplus equipment to the public. They usually have stuff like what you need.

If you can find something that has a HPIB/GPIB bus connector (IEEE-488) then you can connect it to a PC and use program your own interface (the libraries are very simple and very well documented). We did this both in the Army and also at a commercial satellite communications company (ours was to interface with HP spectrum analyzers thru IEEE-488).

SoftDSP Digital Oscilliscope

[francisew]

softdsp.com
I bought one two years ago (around 800$ canadian)
it's pretty good, does 200 MHz / (5 GigaSamplesPerSecond equiv., whatever that means), two channel, USB.
The software isn't great. I don't think there is a linux port... I'm lazy, haven't checked recently.
The actual device is really sweet. If I haven't blown it up in two years, it is pretty solid! (I'm a chemist, and I do things like attach 400V power supplies to it randomly, I'll feel bad if it dies. Or me.)
Good luck!

Re:Cheap PC Oscilloscopes

[unitron]

...type "Cheap PC Oscilloscopes" into the search box and click "Search".

If you google for "Cheap PC Oscilloscopes" with the quote marks included all you get is a link to this article.

PCI based oscope

[Triode]

Many people make them, and they are real oscopes on
a PCI card... but the ones I have used were GaGe...

http://www.gage-applied.com/

Should not be too pricey, and I think they have
educational discounts. They are the best option
I have seen to get a real oscope in a computer, and
the sampling rate and digitization will beat a sound card hands down.

Take a trip down to your local electronics swap...

[Ho Kooshy Fly]

Personally you have to be careful about many of the "PC" based scopes that are out there. They usually suffer from bad analog bandwidth and short memory depth. Also you have to pay particular attention to many of the voltage/impendance limits of "PC" scopes.

Personally taking a trip down to your local electronics swap meet is not a bad idea if you have one nearby. Hear in SiValley there are a few around on the weekends where you get some older Tek/HP *cough* Agilent scopes for pennies on the dollar. Sometimes they need some work but most people are honest about it.

-Ho

Re:MHz vs. MSPS

[Anonymous Coward]

Nyquist is about waveforms. If you don't understand the relation between waveforms and frequencies, then you and I are just going to have to drop this conversation right now.

You need 2x samples to capture a waveform. Mathematically. However, we are in the real world, and 5x is considered generally adequate. Truthfully, most engineers won't let a 40MHz scope anywhere near a 40MHz signal; equipment doesn't tend to perform optimally at its edges.

As to why buy a digiscope for repeating signals: data capture is nice. Especially when you can capture raw sample values and spit them into MATLAB to do some serious data analysis. Also, there's the fact that you can't buy new analog scopes anymore. Finally: most DSOs will also function as a spectrum analyzer. Which is great on your repeating signals.

One to Avoid

[certsoft]

I bought a PC Multiscope from Jameco and found out that it didn't actually sample both channels simultaneously, making it basically worthless. One of the things you want to be able to do is to look at clock and data lines and measure the timing between them, and for that you need simultaneous sampling. Needless to say, that pile went back to Jameco.

I bought a DSO-2100 from Link Instruments and have been very pleased with it. Probably one of the best investments I've ever made.

Optascope

[GarthSweet]

I have an Optascope 81M USB scope from
http://www.optascope.com
I recommend it highly. Nice software and really works well for me. Also only $189.00!

Specs are:
1 Ms/S Maximum Sample Rate (500Ks/s 2 channels)
200 KHz Bandwidth
20Vpp Max Input for CH1 & CH2
8 Bit Vertical Resolution
2 Channel
External Trigger Source
Trigger on Rising or Falling Edge at Any Voltage
Variable Trigger Voltage on DSO channels
10%, 50% and 90% Horizontal Trigger Position
Auto or Normal trigger modes
USB interface

Dupe!

[bishr]

This was covered a few years ago; no surprise that most people forgot. The answers are mostly the same; still a good read though: http://ask.slashdot.org/article.pl?sid=01/12/07/19 1220&mode=thread
-----
Trogdor the Burninator!

cheap 40 MHz scope

[RadioDude]

The most common source for these type of PC-card "instruments" is National Instruments.

If you are teaching about electronics, you would be better off buying a used Tek 475 (or similar) analog scope, you can get a very good one for $300-400. They can learn about the actual circutry, timing, measurement error, etc. without
getting heavily into sampling theory and digitial
signal processing.

If you want the students to learn "the new way" of
electronic instruments, check with National Instruments about used/traded in cards, and software; they may have an educational discount.

There are also some "poor man's" type of scopes made of surplus parts, old TV's, etc.. that you can find in the back pages of Nuts& volts magazine; I don't recommend these if you want the students to learn what they will use in the future, in real-world engineering applications.

Finally, there are mixed-mode instruments that are analog with analog storage, analog with digital storage, analog with digital readouts added, various standard instruments with serial or GPIB interfaces, and s/w from the mfr or 3d party for
control and analysis.

See if you can find on some engineer's shelf a catalog/book from Tek or HP, say, from the 1980s
or 90's, this is about the vintage that will work and be in your price range.

Everything oscillates. Okay, a small exaggeration.

[Futurepower(R)]

Modern circuits are not so well controlled as someone might guess. To have 20 kHz output with little phase distortion, it is necessary to pass more than 200 kHz.

But that's not the issue. ICs allow the design of circuits with bandwidths that are literally physically impossible with discrete components. The fundamental bandwidth limitations of the transistors used in the IC may be 200 MHz or more, or even 1.5 gigahertz. Any small problem can cause a circuit to oscillate at 50 MHz, even if the IC is supposedly limited to far less than this.

All you need for oscillation is gain and some positive feedback. In the real world, circuits try all possible combinations almost instantaneously, and begin oscillating for reasons the designer never foresaw. For example, maybe there is capacitive coupling through the IC packaging, and the output circuit alone is oscillating.

This is only a slight exaggeration: There are 4 steps toward making a new electronic device: 1) Build the circuit. 2) Supply power for the first time. 3) Apply an oscilloscope probe and begin discovering all the reasons the circuit is oscillating when it shouldn't. 4) Then discover all the other reasons the circuit isn't working correctly, if any.

I was never a person who thought that killing people and destroying their property was a good way to resolve social problems, but at one time it was my job to repair the automatic flight control systems of fighter-interceptors in the U.S. Air Force's Air Defense Command.

These aircraft required 250 hours of maintenance per hour of flight. (Aircraft meant to be sold to other nations, also, required 15 hours of maintenance per hour of flight. I've followed the development of weapons systems ever since, and my opinion of what is actually delivered is that it is often fraud, or close to fraud. United States taxpayers: Your assigned duty is to find the money to pay, and to avoid thinking.)

Anyhow, during training flights it was required to pull several g's. Sometimes at high accelerations the electronics would go completely crazy, and all inertial reference would be lost. The only fix for this at the time was to land, regain stability, and take off again. The aircraft that had this problem were therefore not much use for any situation actually requiring defense. Worse, the problem seemed to have nothing to do with any particular aircraft, but seemed random.

One day while trying to make a faulty system work on my bench test mockup, I discovered the reason. Some of the amplifiers that controlled the gyros had high frequency parasitic oscillations at perhaps 100 times the normal frequency of operation. You couldn't see the oscillations with normal equipment because the frequency was too high. I had borrowed an oscilloscope from some co-workers who worked on faster electronics.

The design of the amplifiers was acceptable, but many of the amplifiers had bad solder joints. Those with bad solder joints would oscillate; oscillating amplifiers would amplify at the required low frequency, but had a much smaller dynamic range than amplifiers that were not oscillating. (Yes, that bad solder joints could cause this doesn't make much logical sense, but most parasitic oscillations don't make much logical sense.)

The amplifiers had other defects that caused them to have a high failure rate. Every time an amplifier was pulled from an aircraft for a conventional repair, an amplifier was drawn from stock and put in the waiting aircraft so that the aircraft would be immediately operational. That was the reason the instability problems kept moving from aircraft to aircraft.

I drew a circuit diagram of my test setup, wrote an article about the problem, had a photographer take photos of the test setup, took screen photos of the parasitic oscillations, and sent everything to those who review such things. This had several effects. Someone at Air Defense Command headquarters wrote a letter praising my work. When everyone was reviewed for p

AC coupling isn't a problem

[rsw]

So modulate in hardware and demod to baseband in software.

With a simple analog multipler (for example, the Analog Devices AD834) and e.g. a 5 KHz oscillator, you can AM a band-limited (say, DC-500 Hz) signal, put it in your sound card, then do the demod in software (another multiplication will work).

This will cost you, in total, about $5 (you can get free samples of the AD834 and you'll need some resistors, some caps, a couple op-amps, and some wire) and will give you DC-500Hz through your modulator or 20Hz-24KHz without it. Not too shabby, especially compared with $500.

By the way, if you're going to spend $500ish anyway, why not pick up a Tektronix 2445 or 2465 on EBay? The 2465 has 350MHz bandwidth and is, IMHO, one of the nicest all-around scopes out there.

Re:Like Skee Lo, I wish...

[aastanna]

Sounds to me like the question is from a high school teacher. I seriously doubt spending any more on a scope would be worthwhile, since I'd bet it's primarily going to be used to show students wavy lines, or maybe as a glorified multimeter (he's only buying one, so it's not like the students will be using it regularly). I wish the poster gave a few examples of the types of projects his class works on.

That said, I wish schools would stop blowing their entire budgets on computer labs. I hate to see a school paying for a computer that isn't at least two years old when the money could be going to text books that aren't twenty years old and falling apart.

It would be really nice to see high school electronics courses teaching students how to properly work a scope, but you'd need enough for an entire class. It's amazing how many engineering students get 2 years into a computer/electrical engineering program and don't know how to use a scope to read important parameters from a circuit.

Please... Don't go the PC route for this...

[KC7GR]

For the amount of money you're specifying, you can easily pick up a much better built and far more versatile O-scope in the form of an older Tektronix instrument.

In fact, that price range will easily get you a 475 or 475A, good to 200 or 250MHz, respectively. It will also put you well within reach of a nice Tektronix 7000 series benchtop 'scope, like a 7704 or even a 7904.

No matter what you may hear, the PC was never designed to be an O-scope, and no amount of external hardware, I believe, will ever turn it into anything that can compare, in terms of value for the $$ and quality of construction, with early Tektronix hardware.

I believe it's also EXTREMELY important to teach would-be technicians and engineers that the PC is not the be-all and end-all of test gear. Never has been, never will be. Oh, it can be useful as an instrument CONTROLLER in automated test setups, yes, but it was never intended to replace the functionality of actual made-for-purpose test equipment.

Give your students a real education. Get a real oscilloscope.