Slashdot Mirror

Yeah it seems those simple bar-displays per frequency I was calling "vu-meter" are also called "spectrum analysers", but it's exactly what you indicated on the link. Most old audio equipment does not use a FFT, but a series of filters (check http://www.electronics-lab.com/blog/?p=7057 , theirs is with nixie tubes, but the principle is the same). Modern mp3 players do not need much effort to reproduce a bar-based spectral analysis, because usually the file format is already encoded on the frequency domain.
Btw, when I think spectrum analyzers, I think of this kind of equipment: http://www.tek.com/spectrum-analyzer/rsa5000 .

I don't know about the exact designs that you have in mind, but I wouldn't want to work on anything that goes much above 10 Mhz as tolerances are so much lower and things get so much harder to design and debug.

You shouldn't need a really high sample rate, and usually 20-40 Mhz of bandwidth on 2-4 channels is plenty.

I've used two really nice digital scopes, by Agilent and by Tektronix, while in school. Both companies have entry-level models that are right in your range, around $1,100.

http://www.tek.com/products/oscilloscopes/tds1000_tds2000/
http://www.home.agilent.com/agilent/product.jspx?nid=-34250.884298.00&cc=US&lc=eng

I've also used a high-end Agilent scope that included a 20(?) channel digital logic analyzer and a 4-channel analog scope and ran visualization software on a built in intel PC running windows.
While this was great for some projects, most of the time it was overkill and I much preferred the "simple" Tektronix scope that didn't have to boot up windows before it was ready to go :)

Don't overlook "old" analog scopes. They work just as good as the latest and greates for most uses, and can be found for a lot less.

Before you buy, try to get a better understanding of your actual needs. How many channels are you actually going to capture at any given time? What's the maximum signal frequency that you'll be working with?

However, I will admit that this is certainly beyond the capabilities of most people due to the lack test equipment that is needed to even test the parts found at the swap meet.

True. Although it is easier to get gigahertz test gear than it used to be, the typical 'scope won't go there.

I would like to add a few ideas to the above. I work at a facility that actually uses these things.

We have some TDS5104B's. They're great scopes, don't get me wrong, but the Windows interface might throw you. I won't go through all the downsides of a Windows OS on a scope, because they're pretty much the downsides of the Windows OS on anything, but the upsides are:
* You can run things like Labview and Matlab right on the scope, and there are libraries that let you talk to the scope itself and control it;
* You can remotely operate the scope via standard VNC programs; and
* You can run programs on the scope that you would otherwise need another computer to do.

As an example, I've been able to download new firmware code to a board via an Altera Stand-Alone Programmer program and a USBBlaster, and watch the result on the scope from my office down the hall from the lab. Another engineer rigged the scope, an Ethernet-equipped function generator, and Matlab on the scope to make a homebrew Bode plotter.

That all said, the TDS5000 series is old and not likely to be sold by Tek too much longer. The model we have, the 5104, is no longer available. There are other models.

The DPO4000 series is comparable to the TDS5000s, except they're half as big, don't run Windows, and can decode serial (e.g. SPI, I2C, or UART) data for you (and let you trigger on those serial patterns with an add-on chip). For our next scopes, we're looking at the MSO4000 series, which are DPO4000s with 16 logic inputs as well, so you can see everything on both the analog and digital sides of an ADC, for example.

The DSA8200 is insanely expensive ($150k or so once you buy the probes) and probably not something you want to let students near. If you aren't designing things like 3.125 Gb/s data links (e.g. SATA or XAUI), these are a waste of money.

The thing is... 90% of the time, we don't need anything that fancy. Which is why we got about half our engineers Tektronix TPS2024s. They're small, simple, portable (battery or wall powered) digital scopes, 200 MHz, with 4 isolated channels. Isolated channels are great in that you can use them to look at differential signals without needing a special differential probe or needing to rig two channels together and use the math channel to take the difference (which you can't trigger on). They also have CompactFlash slots which can be used to grab waveform and setting data and copy it to your computer as CSV files. I have mine set to save everything to CF when I press the PRINT button.

For even smaller work, Agilent has some neat two-channel handheld scopes, their U1600A series. I saw some in their demo trailer this week, and the screens were nice and fast, unlike older handheld scopes and scope/meters.

Must haves? What? Besides being able to show a sine, square, and sawtooth waveform, what more are you looking for? Even the legacy oscilloscopes from HP that I've used has had a lot of the modern features that you see on the newer ones today. Sure, the newer ones do certain things more nicely, but there honestly isn't a huge difference, unless you're looking for things like color, USB support for capturing waveforms, super accurate frequency readings, etc... What you need to do is figure out what specifically is going to be needed with whatever projects you're doing.

This is what I use at work. (Specifically the TDS2000B.) I have no complaints with it. I've found this to be intuitive to use. It's simple and robust for what it is. Other people around me have to use the TDS1000B, and really the only difference is the lack of a multicolor display, USB support, and only 2 channels. I haven't had any issues showing our "trained monkeys" (pre-testers who have zero training or education in electronics) on how to use these oscilloscopes.

Depending on the resources in your lab, and its purposes, you might find that a software-defined interface is more flexible for your needs. You can add any sort of interface or processing capabilities you want.
http://sine.ni.com/nips/cds/view/p/lang/en/nid/205615

If it's a research lab, something like the above may be just what you want. If it's a teaching lab, and you want students to have access to real knobs and buttons, then my experience is specifically with Tektronix scopes that I use at work. Again, without knowing your price range, there are a wide range of options out there.

At the low end, the TDS5054B series has an interface likely identical to that of your old scope; they did a reasonable job of replicating the older style of analog interface but added on some processing utilities.
http://www.tek.com/products/oscilloscopes/tds5000b/

I've used a scope the DPO400 series as well. I found it's interface to be rather, well, different at first, with all the options not in the places I'd usually expect them to be with my other Tek scopes. But I eventually grew used to it and found it all perfectly fine - except that the probe connections for some reason don't allow use of our current probes. They work fine on all other scopes, and I see no reason why they molded the plastic on this scope to exclude them.
http://www.tek.com/products/oscilloscopes/dpo4000/

Finally, at the high end, you have something like the DSA8200. This scope runs windows, which you can get to to do some data analysis, but the scope itself is controlled through the Tek application. It looks and behaves like a piece of software; there are buttons on the front for some features, but they are just macro buttons to execute the commands; it's often faster to just use a mouse since the buttons only offer limited functionality.
http://www.tek.com/products/oscilloscopes/dsa8200/

Note how the Tek scopes are all mostly more expensive than the NI scope, with more limited flexiblity in the interface. Again, if I knew what bandwidth you needed or what your budget was or the purpose of your lab, I could give better recommendations.

Depending on the resources in your lab, and its purposes, you might find that a software-defined interface is more flexible for your needs. You can add any sort of interface or processing capabilities you want.
http://sine.ni.com/nips/cds/view/p/lang/en/nid/205615

If it's a research lab, something like the above may be just what you want. If it's a teaching lab, and you want students to have access to real knobs and buttons, then my experience is specifically with Tektronix scopes that I use at work. Again, without knowing your price range, there are a wide range of options out there.

At the low end, the TDS5054B series has an interface likely identical to that of your old scope; they did a reasonable job of replicating the older style of analog interface but added on some processing utilities.
http://www.tek.com/products/oscilloscopes/tds5000b/

I've used a scope the DPO400 series as well. I found it's interface to be rather, well, different at first, with all the options not in the places I'd usually expect them to be with my other Tek scopes. But I eventually grew used to it and found it all perfectly fine - except that the probe connections for some reason don't allow use of our current probes. They work fine on all other scopes, and I see no reason why they molded the plastic on this scope to exclude them.
http://www.tek.com/products/oscilloscopes/dpo4000/

Finally, at the high end, you have something like the DSA8200. This scope runs windows, which you can get to to do some data analysis, but the scope itself is controlled through the Tek application. It looks and behaves like a piece of software; there are buttons on the front for some features, but they are just macro buttons to execute the commands; it's often faster to just use a mouse since the buttons only offer limited functionality.
http://www.tek.com/products/oscilloscopes/dsa8200/

Note how the Tek scopes are all mostly more expensive than the NI scope, with more limited flexiblity in the interface. Again, if I knew what bandwidth you needed or what your budget was or the purpose of your lab, I could give better recommendations.

Depending on the resources in your lab, and its purposes, you might find that a software-defined interface is more flexible for your needs. You can add any sort of interface or processing capabilities you want.
http://sine.ni.com/nips/cds/view/p/lang/en/nid/205615

If it's a research lab, something like the above may be just what you want. If it's a teaching lab, and you want students to have access to real knobs and buttons, then my experience is specifically with Tektronix scopes that I use at work. Again, without knowing your price range, there are a wide range of options out there.

At the low end, the TDS5054B series has an interface likely identical to that of your old scope; they did a reasonable job of replicating the older style of analog interface but added on some processing utilities.
http://www.tek.com/products/oscilloscopes/tds5000b/

I've used a scope the DPO400 series as well. I found it's interface to be rather, well, different at first, with all the options not in the places I'd usually expect them to be with my other Tek scopes. But I eventually grew used to it and found it all perfectly fine - except that the probe connections for some reason don't allow use of our current probes. They work fine on all other scopes, and I see no reason why they molded the plastic on this scope to exclude them.
http://www.tek.com/products/oscilloscopes/dpo4000/

Finally, at the high end, you have something like the DSA8200. This scope runs windows, which you can get to to do some data analysis, but the scope itself is controlled through the Tek application. It looks and behaves like a piece of software; there are buttons on the front for some features, but they are just macro buttons to execute the commands; it's often faster to just use a mouse since the buttons only offer limited functionality.
http://www.tek.com/products/oscilloscopes/dsa8200/

Note how the Tek scopes are all mostly more expensive than the NI scope, with more limited flexiblity in the interface. Again, if I knew what bandwidth you needed or what your budget was or the purpose of your lab, I could give better recommendations.

The hardware unlocking doesn't even require a soldering iron. It could be done a lot more easily with something like this. Much lower probability of bricking your $600 phone, and much less effort required as well.

Bah. HP changed when they shifted their focus from engineering to sales. They used to make some of the world's best scientific equipment, but they sucked the soul from that and spun it off. Likewise some of the best scientific calculators. Now their fortunes ride the tide of ink sales and everybody thinks of them as a PC supplier.

Oh well, at least we've still got Tektronix...

tekkies (love that spelling)

Are those fans of Tektronix or Shatner's TekWar?

1) They are mounted on a PCB

2) A lot of engineering went into ESD protection diodes at the pins?

When you realise that the gate on a modern CMOS transistor is only about 5 atoms thick, and you calculate the energy density of a typical human body model ESD discharge, you see that you can quickly degrade those transistors.

The fact that the factory uses ESD practices throughout the plant, allows for fuck-head know-nothing techs like you to get away with not using a strap and jumping to faulty conclusions.

Thank fuck you're not handling my 12GHz sampling heads! I'd shove them up your ass sideways if I saw you connecting a coax without discharging it first!

I have no experience or contact with oscilloscopes, so would someone please inform me why they cost so damned much, even used (and up to friggin $20K, new!).

Target market: High end 'scopes like a digital storage oscilloscopes are made for a target market of an electronic engineering company, or electronic test centres, not your local (radio) ham or TV repair person. So does a company worry about spending $20K for a piece of equipment to sit on the desk of a $75-150K EE who is desinging something expected to make millions of dollars profit? No.

Oscilloscopes are built to last, the 100MHz analog one I own is older than I am I think.

Tektronic and HP/Agilent build them to not be obsolete, and the replacement cycle is likely about 5 years or longer depending on what electronics market segement your in.

All you need is a big old satellite antenna, a spectrum analyzer, a frequency generator capable of tuning to a stable 6 or 14 GHz (C or Ku) signal, an amplifier capable of amplifying that kind of signal up to about 100 watts, some waveguide... hmmm. It would probably also help to have a fast car for when the FCC arrives 30 minutes later.

This is sort of like saying that flying to the moon is easy - all you need is a big metal tube and some propellant.

Check out this O-scope:

http://www.tek.com/Measurement/Products/catalog/td s7000/eng/index.html

It runs Windows 98! I evaluated it for purchase by the company I work for. It seemed to work well but we chose a Lecroy scope instead because in general I haven't been happy with support for any PC based control system. PC components seem to go into obsolescence far to quickly. O-Scopes should last many years and I am skeptical about getting parts for a PC-based scope at a time when it is 10 years old.

I didn't look into it, but they may be using some 3rd party software to replace the Window's kernal with a RTOS. I ran some control software that did this for WinNT 4.0. I think that it was called Intime.

This control software:
http://www.entivity.com/vlc.htm
replaces the HAL of WinNT with a RTOS. You can crash NT and rip out the HD and the control software still runs, no HMI though. It had a crappy programming environment (flowcharting) but I was sort of impressed with its stability.

• linuxtv
• inria.fr
• dvb.org
• protocol notes
• gilat
• europeonline
• drdish
• helius
• telemann.com
• broadcast.philips