How much scope BW do you need?

[w2aew]

A question I'm often asked is - how much BW do I need get with my oscilloscope.  I usually answer with - get as much as you can afford!  Even low frequency circuits can do some wacky high frequency things.  Check out my latest video for an example...

[BravoV]

Excellent tutorial video w2aew, thanks.

I used to have a 60Mhz analog, when upgraded to Tek 2465B paired with P6137 probes, this thing is really an eyes opener, fixed a lot of circuits that I built that behaved strangely because of spurious high freq oscillation. The 60 Mhz simply won't cut it anymore, and I always thought that playing with audio level circuit never need a high freq scope, I was wrong. 

Really drolling seeing that fine piece of Tek 2467 of yours. 

Btw, apart from the good thing you've shown here, the bad news is this kinda video makes it even harder to hunt down a good price of an old good analog scope then before.... j/k.

PS : I think this video should be posted at your stickied thread as well rather than buried deep under other threads.

[EEVblog]

Nice find, and excellent example.

Dave.

[firewalker]

Nice one!

Alexander.

[w2aew]

Excellent tutorial video w2aew, thanks.

I used to have a 60Mhz analog, when upgraded to Tek 2465B paired with P6137 probes, this thing is really an eyes opener, fixed a lot of circuits that I built that behaved strangely because of spurious high freq oscillation. The 60 Mhz simply won't cut it anymore, and I always thought that playing with audio level circuit never need a high freq scope, I was wrong. 

Really drolling seeing that fine piece of Tek 2467 of yours. 

Btw, apart from the good thing you've shown here, the bad news is this kinda video makes it even harder to hunt down a good price of an old good analog scope then before.... j/k.

PS : I think this video should be posted at your stickied thread as well rather than buried deep under other threads.

Yes - I'm lucky to have the 2467.  It would be VERY tough to see the 200MHz oscillation using the dual delayed timebase with most other analog scopes - they don't have the phosphor writing rate of the 2467. 

Oh, and good idea - I'll post a reply to the sticky thread...

[saturation]

A great video.

If we go by the rule of buying as much as we can afford, we'd all be shooting for an Agilent 9000x class scope in the $300,000 range.  But, once we enter the 33GHz arena, what are we not seeing above it that a faster LeCroy can see at 45GHz?  And the list goes on, not to mention the cost of the scopes.

In RF region, subtleties of probe design and placement matter.  Where the parasitic oscillations truly there or induced by the act of examining it with a probe, by introducing the probes impedance and capacitance?  I don't know.  One way to find out is to check the distortion in the collector output waveform, did it improve with the base resistor? or one can double check the base input with an active probe.

So the larger question is, knowing the more you measure the more you see, when do you stop?

I think a more lasting rule is we need to be aware that if we design for X and get X-1 or X+Y, that something is admist and needs to be found, explained or corrected.  We shouldn't rely on instruments to find faults by blindingly probing without some conception of how the fault can be formed; so one suspect the faults first and then confirm it with an instrument adequate for what we suspect exists.

Likewise, in complex designs if it does what it was designed to do under the conditions of operation,  there is probably nothing more to do, even if there is an unintended oscillation somewhere, because if it is significant it will cause its output to not be the X we intended.  So in the example given, if the phase shift oscillator is supposed to produce a low distortion sine wave, which the output clearly wasn't, then it justifies looking for its cause.  This simpler rule will last you a lifetime, is not dependent on equipment, and will work on just about anything you work on and will help one choose equipment cost effectively.

[Mechatrommer]

3KHz circuit -> 200MHz oscillation (multiple by a factor of 66,667). so say i'm building 1MHz circuit, so i should expect probably 67GHz spurious and get a 670GSps scope / 67GHz spectrum analyzer? shoot me! with 50MHz scope, i'm only eligible to design a 0.75KHz circuit which is 750Hz. think i better off with "mains" only job

[IanB]

3KHz circuit -> 200MHz oscillation (multiple by a factor of 66,667). so say i'm building 1MHz circuit, so i should expect probably 67GHz spurious and get a 670GSps scope / 67GHz spectrum analyzer? shoot me! with 50MHz scope, i'm only eligible to design a 0.75KHz circuit which is 750Hz. think i better off with "mains" only job

Even if you can't see the oscillation, you may be able to see the effects of the oscillation. For example irregular waveforms, poor regulation, lack of stability, excess power consumption, etc. Seeing the actual waveform is only proof of a suspicion, not a required analytical tool. One must always develop a sixth sense for problems and learn to get the best out of the tools available.

[w2aew]

A great video.

If we go by the rule of buying as much as we can afford, we'd all be shooting for an Agilent 9000x class scope in the $300,000 range.  But, once we enter the 33GHz arena, what are we not seeing above it that a faster LeCroy can see at 45GHz?  And the list goes on, not to mention the cost of the scopes.

In RF region, subtleties of probe design and placement matter.  Where the parasitic oscillations truly there or induced by the act of examining it with a probe, by introducing the probes impedance and capacitance?  I don't know.  One way to find out is to check the distortion in the collector output waveform, did it improve with the base resistor? or one can double check the base input with an active probe.

So the larger question is, knowing the more you measure the more you see, when do you stop?

I think a more lasting rule is we need to be aware that if we design for X and get X-1 or X+Y, that something is admist and needs to be found, explained or corrected.  We shouldn't rely on instruments to find faults by blindingly probing without some conception of how the fault can be formed; so one suspect the faults first and then confirm it with an instrument adequate for what we suspect exists.

Likewise, in complex designs if it does what it was designed to do under the conditions of operation,  there is probably nothing more to do, even if there is an unintended oscillation somewhere, because if it is significant it will cause its output to not be the X we intended.  So in the example given, if the phase shift oscillator is supposed to produce a low distortion sine wave, which the output clearly wasn't, then it justifies looking for its cause.  This simpler rule will last you a lifetime, is not dependent on equipment, and will work on just about anything you work on and will help one choose equipment cost effectively.

Of course, my "general answer" to getting as much BW as you can afford, was really somewhat in jest.  This is usually the advice that I'll give to a fellow ham or hobbiest, not to engineers looking for a scope for their job.  In those cases, there is a LOT more discussion on the application and measurement requirements to help determine whether they need a DPO73304D 33GHz scope, or maybe a MSO72004 20GHz mixed signal scope, or maybe a MDO4104-6 mixed domain scope with spectrum analyzer, etc.  It's the ham or hobbiest that will typically be looking at a low cost used analog scope or maybe an inexpensive digital scope - it is these guys that I would typically advise getting a little more BW than they think they need (all other things being equal).

You are absolutely correct regarding the subtleties of RF design.  I purposely picked a really fast switching transistor to throw in the lousy solderless breadboard in order to "create" the oscillation problem - to point out that sometimes these things happen without you expecting it.  Some folks asked if the oscillation was induced by the probe capacitance when I connected it to the base of the transistor - just as you did.  I did another video that shows the oscillation on the collector, with nothing attached to the base.  The same video also shows some setup considerations for looking at this with a digital scope as well.  Here's the 2nd video if you are interested:


As you said, there's no substitute for common sense.  Understanding your circuit and what can go wrong, as well as understanding the capabilities and limitations of your equipment so you don't get fooled, are equally important.  Its always nice to have a few ways of looking at a given measurement.

[vk6zgo]

This would be a useful way of checking all the random unmarked transistors lying about,to see how fast they are!

Of course,you could check to see if your audio oscillators are producing spurious HF/VHF outputs with an Absorption Wavemeter,or a Grid Dip Oscillator switched to the Wavemeter position,or even a Spectrum Analyser,with a pickup loop.
All of these methods should satisfy the purists,but I'm quite happy with the 350MHz Oscilloscope!

VK6ZGO

[w2aew]

This would be a useful way of checking all the random unmarked transistors lying about,to see how fast they are!

Of course,you could check to see if your audio oscillators are producing spurious HF/VHF outputs with an Absorption Wavemeter,or a Grid Dip Oscillator switched to the Wavemeter position,or even a Spectrum Analyser,with a pickup loop.
All of these methods should satisfy the purists,but I'm quite happy with the 350MHz Oscilloscope!

VK6ZGO

You're right!  I did check for the oscillation with a spectrum analyzer with a near-field pickup loop too - mainly to satisfy myself that I wasn't inducing the oscillation with the scope probe capacitance.  The wavemeter and GDO would require some tedious tuning, but would work fine.

aw

[saturation]

Thanks w2aew, it clarifies a lot.  It would be good for those unfamiliar to see a full point of view as you wrote below.  That follow up video is superb as was the first.  You make great, quality videos, I hope you continue to make more.   There are a lot of videos on youtube and elsewhere, but quality is rare.

Of course, my "general answer" to getting as much BW as you can afford, was really somewhat in jest. 
.
.

[BravoV]

w2aew,

Sorry for going abit offtopic, I noticed at your 2647 when in delay time base mode, your zoomed signal is not at the center horizontally as the main signal above it.

My 2645B is also have this same position when the delay time base switched on, the zoomed part is shifted to the right side exactly like yours.

Is this normal even for a fully calibrated Tek 246x series ?

Here the capture image from your video to show you what I'm talking about, the yellow arrow showing the zoomed part is not at the center position horizontally leaving a bit of unused area at the area pointed by red arrow.

[w2aew]

w2aew,

Sorry for going abit offtopic, I noticed at your 2647 when in delay time base mode, your zoomed signal is not at the center horizontally as the main signal above it.

My 2645B is also have this same position when the delay time base switched on, the zoomed part is shifted to the right side exactly like yours.

Is this normal even for a fully calibrated Tek 246x series ?

Here the capture image from your video to show you what I'm talking about, the yellow arrow showing the zoomed part is not at the center position horizontally leaving a bit of unused area at the area pointed by red arrow.

Good question.  The reason for this is that I have the delayed timebase running extremely fast - 5ns/div as compared to the main timebase which is running 20,000x slower.  When the scope switches from the A sweep to the B sweep, it blanks the trace so that you don't see the retrace.  When the B sweep starts, the trace is unblanked.  The blanking circuit is not quite as fast as the sweep, so you're seeing the the response of the un-blanking of the beam.  At slower timebase settings (for B), you wouldn't see the "shift".  It really isn't a shift, it's just that the beam isn't up to full brightness at the start of the very fast B sweep.

[BravoV]

Good question.  The reason for this is that I have the delayed timebase running extremely fast - 5ns/div as compared to the main timebase which is running 20,000x slower.  When the scope switches from the A sweep to the B sweep, it blanks the trace so that you don't see the retrace.  When the B sweep starts, the trace is unblanked.  The blanking circuit is not quite as fast as the sweep, so you're seeing the the response of the un-blanking of the beam.  At slower timebase settings (for B), you wouldn't see the "shift".  It really isn't a shift, it's just that the beam isn't up to full brightness at the start of the very fast B sweep.

Pheeww.. what a relieve, this explains a lot, cause I've been wondering all this time why my B sweep at high speed ns level is dimmer and shifted right while at slower rate it looks better, thank you very-very much.

With that MCP crt, your 2647 is definitely looks a lot better than my 2645B using the plain jane crt, especially when fiddling at ns level, that piece of beauty is really one of the finest portable high speed analog scope ever created, again.. I'm still drolling when everytime I look at it. ha..ha..

[w2aew]

With that MCP crt, your 2647 is definitely looks a lot better than my 2645B using the plain jane crt, especially when fiddling at ns level, that piece of beauty is really one of the finest portable high speed analog scope ever created, again.. I'm still drolling when everytime I look at it. ha..ha..

You are absolutely right - the 2467 an absolutely fantastic portable scope with the extremely fast writing rate of the microchannel plate CRT, it was (and still is) a gem.  Of course, the 2465B is a great one too - but the CRT in the '67 puts in another class.  I'm very lucky to have it - I picked it up at the Dayton Hamvention about 10 years ago.

[chscholz]

LeCroy is now at 60 GHz.

http://www.lecroy.com/labmaster/labmaster10zi/

A great video.

[...] But, once we enter the 33GHz arena, what are we not seeing above it that a faster LeCroy can see at 45GHz?  And the list goes on, not to mention the cost of the scopes.

[...]

[Gridstop]

w2aew, thanks again for your videos. When I got my 7104 I really had no idea what a delaying timebase was, but your clips cleared it right up. It's amazing that on the 7104 I can have both the overview of the calibration signal, as well as a view so fast that the 'rising edge' is virtually flat.

The only trouble I have is how touchy the delayed time adjustment is, even with the multi-turn pot. But I guess when you're zoomed in 10,000 times or so on the second trace, how far does each degree of a 10 turn pot take you...

[w2aew]

w2aew, thanks again for your videos. When I got my 7104 I really had no idea what a delaying timebase was, but your clips cleared it right up. It's amazing that on the 7104 I can have both the overview of the calibration signal, as well as a view so fast that the 'rising edge' is virtually flat.

The only trouble I have is how touchy the delayed time adjustment is, even with the multi-turn pot. But I guess when you're zoomed in 10,000 times or so on the second trace, how far does each degree of a 10 turn pot take you...

Congrats on the 7104 - another great scope with the MCP CRT.  That CRT really spoils you compared to ordinary CRTs.  I'm glad that some of my videos helped you out!

[slateraptor]

Wow. This thread has a wealth of knowledge! Big thanks to w2aew, saturation for the insightful constructive criticism, BravoV for asking the beam sweep question, and vk6zgo for the non-invasive detection methods (I recall Bob Pease writing about a Heathkit grid dip meter he used for tracking down high-freq oscillations left unattended by other engineers that interfered with his troubleshooting).