Recommended oscilloscope features?

[GARTHWILSON]

Quote:

Would you use phase 2 as the external trigger without wasting the channel to display the phase 2 waveform? or how would a feature like this be used?

Ok, here's an example. On an SPI or Microwire syncrhonous-serial interface, let's say you want to use the two traces to watch clock and MOSI (data in to the slave device), but you want to start the screen when the device's select line goes true. The edge on the select line is not very interesting and you've probably aready seen it anyway, so there's no need to waste a trace on it. Just trigger on it to see the first X number of bytes of a transaction on the two traces you can see and want to see. The select line you're triggering on serves as a marker for the beginning of each transaction.

[GARTHWILSON]

Quote:

There are plenty of cheap probes from Hong Kong on ebay - but there must be room for doubt. Has anyone any experience?

Do they guarantee a minimum level of performance, including up to 400MHz or whatever you're interested in? (The probes should probably be good for more than speed of the scope, to make sure they don't limit the performance.) New ones can easily be $50 or $100 each, or more. This 400MHz one from Probe Master is $175 each.

I seem to remember my pair of 40MHz probes for my 20MHz oscilloscope was between $40 and $50 for the pair, who knows how many years ago.

[fachat]

ElEctric_EyE wrote:

What are you looking to do exactly?, if I may be so bold. For sure Tektronics is a good investment! Looks all digital. I'm jealous.. :cry:

Well, with my electronics development I am getting into the limits of my old Tektronix 531A... It's also not in a good shape anymore. The trigger mechanism isn't really working anymore, even on digital (TTL) signals, so I can only do time-based measurements. And the time base is not calibrated anymore anyway. Lastly it uses up a LOT of space in my room... [EDIT: not to forget the occasionally puff of smoke coming out of it - at least once so far that I know of...]

So far this hasn't stopped me from using it and getting at least some good qualitative results. But my new 65816 card has a 50MHz oscillator on it, and even though none of the signals actually runs at that speed, I just can't see that signal on my scope. And 10-12.5 MHz static RAM access speed isn't too far away here. And I am still looking forward to finally fix these longstanding problems that make my DRAM-based video card so delicate to handle (depending on chip speed, phase of the moon etc). But for this I also need a decent triggering mechanism.

And as I think of going even further not only with CPLD with FPGA based developments, I thought that it's now time for a new scope. And as I believe in good tools, as they just make your life so much easier, it has to be a good one :-)

André

[ElEctric_EyE]

Some of the new Tektronix 200MHz+ probes are going for more than $450US. Is this worth it?

The 'scope I bought on ebay had included what looks like 2 probes that were originally included with the scope. They're not being sold anymore, and since the scope is "ancient" a price comparison is impossible. And tech back then might not be up to par now...

My question is, does the Nyquist theory (2x oversample) apply to 'scope probe's too? Is it worth upgrading probes?

[BigDumbDinosaur]

fachat wrote:

[EDIT: not to forget the occasionally puff of smoke coming out of it - at least once so far that I know of...]

Just open the windows and let the smoke out of the room. Also, think of smoke as a pilot light. If you see smoke it means the power is on.

Knowing that you are in Germany and not knowing what sort of a budget you have to get a new scope, I can't really offer anything other than get the best you can afford. With what you want to do, a 100 MHz unit would, in my opinion, be the minimum to shoot for. As Garth said, probes are equally important, so I guess it becomes a tug-of-war between being able to afford a high-end scope and also afford the high-end probes needed to make the high-end scope produce high-end results.

[GARTHWILSON]

Quote:

My question is, does the Nyquist theory (2x oversample) apply to 'scope probe's too? Is it worth upgrading probes?

A DSO's sampling does not happen in the probes themselves [Edit: It seems that it does in the ultrafast scopes, beyond half a GHz or so; does anyone know?], so I suppose it's just a matter of remembering that the specified maximum frequency is usually at -3dB and the phase shift that goes with it (beyond just the time delay that the length of cable produces which is not a problem), and deciding whether you want to put up with that added to the scope's high-end drop-off or if you want to spend more to get the probes' limit up higher and out of the way.

My 20MHz scope with 40MHz (IIRC) probes does actually show a 70MHz signal, but very attenuated, and of course a square wave looks like a sine wave. IOW, the probes don't have a sharp cutoff frequency or alias all the frequencies that are beyond a certain point. But although they'll show that signal, it's worthless for anything but confirming that there is a signal and seeing the approximate period. (Actually I use my 512MHz frequency counter for the period or frequency though, which interestingly enough can also go beyond 512MHz, in spite of the fact that I originally thought that was a limit on the digital counter registers.)

[ElEctric_EyE]

1 last question Garth, because now I know my probes are inadequate trying to measure an 80MHz oscillator, (even though the scope itself can measure 200MHz signals), and you seem to be the most knowledgeable and equipped in this area.

I see an 80MHz sine wave output from a 1/2size 5V powered square wave oscillator. I'm measuring p-pV @+2.5V. Can I rely on that peak to peak reading? I'm using this oscillator to power 3.3V hardware because of this reading. Can I trust it? TIA

[GARTHWILSON]

Quote:

1 last question Garth, because now I know my probes are inadequate trying to measure an 80MHz oscillator, (even though the scope itself can measure 200MHz signals), and you seem to be the most knowledgeable and equipped in this area.

Well, I don't know that I'm necessarily the most knowledgeable, but working in applications engineering at a UHF power transistor manufacturer in the mid-80's was good hands-on education in the related parasitic elements, transmission-line effects, etc.. I wish I were good at LaPlace transforms, but in my minor digital work just having a good feel for what's going on up there seems almost as good as being able to model and calculate it all. At least I got cozy with the Smith chart after being introduced to it in amateur radio in the 1970's.

Quote:

I see an 80MHz sine wave output from a 1/2size 5V powered square wave oscillator. I'm measuring p-pV @+2.5V. Can I rely on that peak to peak reading? I'm using this oscillator to power 3.3V hardware because of this reading. Can I trust it?

Without seeing it, I could guess that if it looks somewhat square it's probably a fairly good representation for this application where you don't really need much accuracy; but if it doesn't, you wouldn't necessarily know if it's because of the probe or not. If you fed the oscillator signal through a 74ACT04, both with excellent power supply bypassing (leaded monolithic ceramic capacitors won't be good enough— you'd need power and ground planes and chip capacitors used strategically), and the shortest practical connections between them, and no load on the ACT output, should give you an output that swings to within 0.1V of the rails, with good squareness. That should be a pretty good check on how well the probe is doing, putting it directly on the 74ACT04's output pin.

And of course when doing the high-frequency or high-speed measurements, you always use the x10 position, never the the x1 position. The probe is never good for anywhere near its maximum speed in the x1 position.

I wish I could give you a better answer.

[ElEctric_EyE]

This is a lesson learned. Don't always trust what you see on the 'scope, especially if you know your probes are not up to par which I personally know now thanks to this thread.
I trust the oscillator datasheet more, ( http://www.ecsxtal.com/store/pdf/ecs_2200.pdf ), which says 4.5V p-p. It also says it runs @3.3V optional, but no specs. That's another topic not worth writing about....

Just FYI, at least my scope measured the period correct. Thanks Garth!

[ElEctric_EyE]

ElEctric_EyE wrote:

...My question is, does the Nyquist theory (2x oversample) apply to 'scope probe's too? Is it worth upgrading probes?

I meant if one had a 'scope stating 200MHz, would the probes have to be certified @400MHz in order to properly see the 200MHz square wave signal? or did someone say 10X?!

[GARTHWILSON]

Quote:

I meant if one had a 'scope stating 200MHz, would the probes have to be certified @400MHz in order to properly see the 200MHz square wave signal? or did someone say 10X?!

The "10x" setting on the probe has to do with the voltage; ie, 1V will show as 0.1V on the screen. However, a square wave has all the odd harmonics (and a little bit of the even harmonics if the duty cycle is not 50%), with amplitude decreasing as you get into higher harmonics. If you have a 20MHz square wave, and you took out all the frequency content above 20MHz, you'd see only a sine wave. If you left the 3rd harmonic in and took out all the frequency content above 60MHz (the 3rd harmonic) without affecting the amplitude or phase of 60MHz and 20MHz, it would look like it's making an effort to be a square wave, but a pretty lousy one. If you could take out everything above 100MHz (the 5th harmonic) without affecting the amplitude or phase of the lower frequency components, it would improve further, but it's not going to look really square until you get to at least 180 or 220MHz (9th or 11th harmonic). So yes, if you want to get a good representation of a square wave at a particular frequency, the oscilloscope should be good to at least ten times the fundamental.

https://en.wikipedia.org/wiki/Square_wa ... e_wave.gif is an animation of the additive synthesis of a square wave with an increasing number of harmonics

[ElEctric_EyE]

Excellent, I understand. Thanks for taking the time to explain that. Similar principles to audio and waveform generation.

I looked up the probes that came with my scope and found they were 40MHz probes not even made for the scope! I had assumed they were OEM for the 2440. Forgive me, I am abit slow... The correct Tektronix 400MHz probes go for ~$150ea. used which isn't too bad. I'll be ordering them in a few weeks.

[bogax]

GARTHWILSON wrote:

The "10x" setting on the probe has to do with the voltage; ie, 1V will show as 0.1V on the screen. However, a square wave has all the odd harmonics (and a little bit of the even harmonics if the duty cycle is not 50%), with amplitude decreasing as you get into higher harmonics. If you have a 20MHz square wave, and you took out all the frequency content above 20MHz, you'd see only a sine wave. If you left the 3rd harmonic in and took out all the frequency content above 60MHz (the 3rd harmonic) without affecting the amplitude or phase of 60MHz and 20MHz, it would look like it's making an effort to be a square wave, but a pretty lousy one. If you could take out everything above 100MHz (the 5th harmonic) without affecting the amplitude or phase of the lower frequency components, it would improve further, but it's not going to look really square until you get to at least 180 or 220MHz (9th or 11th harmonic). So yes, if you want to get a good representation of a square wave at a particular frequency, the oscilloscope should be good to at least ten times the fundamental.

A bit of a quibble

That's a fine explanation except you leave out the most salient point

The 10x setting presents less capacitance to the ciruit. A byproduct of
doing that (in a simple way) is that it presents 1/10th the voltage to the
scope.

[GARTHWILSON]

Actually there's a lot more to it than just reducing the capacitive load on the circuit, but that wasn't really what I wanted to expand on. But yes, I probably should have said so.

[bogax]

GARTHWILSON wrote:

Actually there's a lot more to it than just reducing the capacitive load on the circuit, but that wasn't really what I wanted to expand on. But yes, I probably should have said so.

Fair enough.

My point was just that it goes along with (your discussion of) preserving
the higher frequencies.