New Analog Scopes?

[PA4TIM]

The fundamental frequency fo, is called the first harmonic, so here f1, is the second Harmonic, so f4 is the 5th harmonic like Gunb wrote.

But I'm not sure about the f1 ect notation, most times I read fo as fundamental, sometimes f0 ( f zero). As far as I know it is fo, if it is fo then f1 ect seems not to be right.
What is the correct notation ?

[madshaman]

Madshaman:
Forget building a 10 GHz (analog crt) scope. There has been a 1GHz realtime Tek, the CRTs for those are extreme and very rare. You will not find a CRT that does 10 GHz without sampling. An analog sample scope only needs a 10 to 20 MHz baseunit.
Try building a 10 MHz scope first, including triggering, several timebases and calibrated V/div input section. Then the right delay network, low noise amplifiers flat over 10 MHz and the high voltage part, 100 MHz scopes go over 10kV for the CRT, and defelection voltages go up too because they are related. Vertical and horizontal amplifiers from analog scopes are state of the art designs and even today used as examples in modern analog design literature. ( i have two books where the writers use Tektronix amplifiers from the 60/70's  for this)

Yeah, soon as the idea occurred I realised that the CRT would be the limiting factor, so 1Ghz CRT is as high as was ever produced?  Which books do you have or could recommend for ground up oscilloscope design?  I think I would like to try and build one, do you think there's merit in learning/building an analog scope first then move to ADC later?  Starting with a lower bandwidth and getting everything right does sound like the best approach to start.

[Gunb]

your statement is a little bit misleading, something like "5. harmonic" could be read as 5th harmonic,
but here it is 9th harmonic you taking about. It is easier, even in the 50% duty case, to call them what they are,
multiplied fundamental frequency.

Agreed. You're right.

[tinhead]

The fundamental frequency fo, is called the first harmonic, so here f1, is the second Harmonic, so f4 is the 5th harmonic like Gunb

let's get through the org. posting:

f0 = 7MHz                      (1. harmonic) <- not exactly,f0 is fundamental, f1 is 1 harmonic - yes, they same, i know it.
f1 = 3*7MHz = 21MHz    (2. harmonic) <- wrong, f1 is 1st harmonic, 3 time the fundamental is 3rd harmonic and not 2.harmonic
f2 = 5*7MHz = 35MHz    (3. harmonic) <- wrong, f2 is 2nd harmonic, 5 time the fundamental is 5th harmonic and not 3.harmonic
f3 = 7*7MHz = 49MHz    (4. harmonic) <- wrong, f3 is 3rd harmonic, 7 time the fundamental is 7th harmonic and not 4.harmonic
f4 = 9*7MHz = 63MHz    (5. harmonic) <- wrong, f4 is 4th harmonic, 8 time the fundamental is 9th harmonic and not 5.harmonic

but i think it should be as following:

f0 = 7MHz or f1= 7MHz   (1. harmonic or fundamental)
f3 = 3*7MHz = 21MHz    (3. harmonic)
f5 = 5*7MHz = 35MHz    (5. harmonic)
f7 = 7*7MHz = 49MHz    (7. harmonic)
f9 = 9*7MHz = 63MHz    (9. harmonic)

We have now 5 components of the waveform but not "5. harmonics".

I don't like wiki as reference, but yeah:
http://en.wikipedia.org/wiki/Fundamental_frequency

and here something about harmonics:
http://www.allaboutcircuits.com/vol_2/chpt_7/2.html

and a nice video as well (for those who like "visual" things)

[nctnico]

The whole purpose of an oscilloscope is that it should produce a pretty picture humans can recognise as a waveform.

If you sample at twice the bandwidth you have all the information in the signal up to the specified bandwidth (the sharpness of the filter determines the actual bandwidth). The downside is that you'll need math intesive algorithms like polynomal approximation to show a signal instead of some seemingly random dots. Another problem is that a very sharp filter causes ringing in the passband so the frequency response isn't flat. Again this could be compensated in software at the expense of processing power. An easy way out is to sample a signal at a samplerate >=5 times the bandwidth (where the samplerate can also be achieved by using time equivalent sampling). This way interpolation is easy and the input filter doesn't need to be sharp which makes it easier to build and gives a flatter (Gaussian) frequency response.

[tinhead]

flatter (Gaussian) frequency response.

Gaussian frequency response is never flat 

To get flat response you need Butterworth- / Chebyshev-like filter, not Gaussian.

[alm]

Yeah, soon as the idea occurred I realised that the CRT would be the limiting factor, so 1Ghz CRT is as high as was ever produced?  Which books do you have or could recommend for ground up oscilloscope design?  I think I would like to try and build one, do you think there's merit in learning/building an analog scope first then move to ADC later?  Starting with a lower bandwidth and getting everything right does sound like the best approach to start.

Yes, I'm pretty sure 1 GHz was the highest, at least in a western commercial design. There may have been some wacky Soviet design with a CRT that was a few meters long that went higher. There have been some transient digitizers (which use scan converter tubes) that went higher, but these are not scopes in a traditional sense (although they used a CRT).

The most in-depth books on scope designs are probably the Tektronix Circuit Concepts book. These were from the sixties, but went into a lot of detail. Later publications are much more superficial. Just compare the current 'ABCs of probes' appnote with 'oscilloscope probe circuits'. Many of them are available here. There also used to be a seller on eBay selling a CD with all of them. Some of the circuits still apply, not all of course. I'm guessing the trigger circuits they show involve tunnel diodes, for example. Many ideas will still be relevant, especially for an analog scope.

An analog scope will involve some extra complexity (the CRT itself, high voltage supply, driving the plates). Not sure how this is any easier than hooking it up to an ADC. Could still be educative, however. I would probably get a CRT from an old scope, preferably one with documentation. Manufacturing these yourself will be a huge project, and using a TV/computer CRT will seriously limit the bandwidth.

[PA4TIM]

The tektronix circuit concepts books are great. I have a hole sersies of them, they were still unread when I bought them. Too bad I have the TDR version only as pdf.

Seibt, handbook of oscilloscope technology is also a nice reading and Tek had a book that was called troubleshooting your scope ( or something like that, also very educationl)
Both are downloadable from Kurts Tek Wiki.

Some time agoo I got a rare gemanium transistor missing for my 1S1 1GHz sample plugin. I placed it today instead of the silicium I used and could not wait to try. The 1S1 performed less as my Philips 1 GHz sample scope but with this transistor it now is great. Nice sharp traces, before this I had to fiddle the trigger knob to find the sweetspot, now I had a trace at once and could use the trigger to stabalize it. So i'm very happy. ( and now time to repair my second 1S1)

Used a very fast Tek pulser ( > 100 MHz) to test it, nice clean Pulses. Then I connected it to my DSO ( 350 MHz Hameg) to play with max sample rates, max wfu/s, peakdetect ect.
Very educational, i managed to get "fantasy" traces. One moment my pulse had several spikes growing from the edges of the pulse. ( 105 MHz pulse at something like 25 Gsa/s equivalent sample time ). Very funny but I was amazed upto what speed the Hameg still gave a rather recognisable  shape of the pulse. Even at 100MHz you could still see it was a squarewave and not a sinewave.

[robrenz]

I often find myself switching to an analog scope when the resolution of a DSO proves insufficient. A recent example was a low level burst of RF oscillation over part of the cycle of a repetitive low frequency signal. Completely invisible on the DSO (buried in the "digitization" noise) but clear as bell on the old Tek.

Yeah, like this:

I have a video uploaded and queued that shows how you can easily miss stuff on an analog scope that gets shown easily on a digital scope.
(common mode noise is the example I used)
Whether or not you want that stuff shown is another argument entirely, but it does show that analog scopes have disadvantages when it comes to this sort of stuff too.

I watched the video, It was excellent but I didn't see what would be missed on an analog scope?  This would be an great time to do a contrasting follow up video about why a high gain differential probe/amplifier would be nice to measure the output ripple.

[madshaman]

Yes, I'm pretty sure 1 GHz was the highest, at least in a western commercial design. There may have been some wacky Soviet design with a CRT that was a few meters long that went higher. There have been some transient digitizers (which use scan converter tubes) that went higher, but these are not scopes in a traditional sense (although they used a CRT).

This kind of technology interests me too :-)

The most in-depth books on scope designs are probably the Tektronix Circuit Concepts book. These were from the sixties, but went into a lot of detail. Later publications are much more superficial. Just compare the current 'ABCs of probes' appnote with 'oscilloscope probe circuits'. Many of them are available here. There also used to be a seller on eBay selling a CD with all of them. Some of the circuits still apply, not all of course. I'm guessing the trigger circuits they show involve tunnel diodes, for example. Many ideas will still be relevant, especially for an analog scope.

Oh thanks!  That's an awesome link, reading through these will be better than sex; I mean that quite literally.

An analog scope will involve some extra complexity (the CRT itself, high voltage supply, driving the plates). Not sure how this is any easier than hooking it up to an ADC. Could still be educative, however. I would probably get a CRT from an old scope, preferably one with documentation. Manufacturing these yourself will be a huge project, and using a TV/computer CRT will seriously limit the bandwidth.

Well, I do have a vacuum system and know a glass blower... but no, seriously, not quite ready to add CRT construction to the table, got enough on my plate at the moment .

It's settled, I'm going to go for it, starting with buying a couple used oscilloscope CRTs, hopefully the same make (I usually get at least two of a new thing I'm going to try working with, chances are good I'll kill one before I know what I'm doing).

The goal will be a 10Mhz scope, that should probably keep me busy.

[vk6zgo]

I often find myself switching to an analog scope when the resolution of a DSO proves insufficient. A recent example was a low level burst of RF oscillation over part of the cycle of a repetitive low frequency signal. Completely invisible on the DSO (buried in the "digitization" noise) but clear as bell on the old Tek.

Yeah, like this:

I have a video uploaded and queued that shows how you can easily miss stuff on an analog scope that gets shown easily on a digital scope.
(common mode noise is the example I used)
Whether or not you want that stuff shown is another argument entirely, but it does show that analog scopes have disadvantages when it comes to this sort of stuff too.

I watched the video, It was excellent but I didn't see what would be missed on an analog scope?  This would be an great time to do a contrasting follow up video about why a high gain differential probe/amplifier would be nice to measure the output ripple.

I'm sorry,Dave,but I didn't really see anything in that video where a digital 'scope was necessary,either.

The interfering signal was repetitive,& an analog instrument would have no problem seeing it.

Even if it just appeared as a thickening of the trace,the first thing I would do would be have a look at 5ms/div to see if it is 50Hz or 100Hz hum,then look at higher frequencies.----maybe a bit slower,but still easy ehough.

Actually,once you know it isn't hum,you don't really need to accurately identify its frequency,just find the source.

Or are we not  talking about the same video?


Sorry again,Dave,I just saw your new video,haven't watched it yet

[nctnico]

The fundamental frequency fo, is called the first harmonic, so here f1, is the second Harmonic, so f4 is the 5th harmonic like Gunb

let's get through the org. posting:
but i think it should be as following:

f0 = 7MHz or f1= 7MHz   (1. harmonic or fundamental)
f3 = 3*7MHz = 21MHz    (3. harmonic)
f5 = 5*7MHz = 35MHz    (5. harmonic)
f7 = 7*7MHz = 49MHz    (7. harmonic)
f9 = 9*7MHz = 63MHz    (9. harmonic)

We have now 5 components of the waveform but not "5. harmonics".
and a nice video as well (for those who like "visual" things)

You are right. Harmonics are numbered as multiples of the fundamental frequency. In the video you can clearly see that a real square wave has even harmonics as well (due to the finite rise/fall time).

[jahonen]

You are right. Harmonics are numbered as multiples of the fundamental frequency. In the video you can clearly see that a real square wave has even harmonics as well (due to the finite rise/fall time).

Actually, even harmonics are produced by some sort of asymmetry of the waveform (for example, duty cycle not exactly 50%), not due to finite rise/fall. Or, of course if rise/fall times are different, that is also asymmetry which qualifies. Finite rise/fall times cause the envelope of the harmonics to fall faster (40 dB/dec) after some frequency determined by the rise/fall time.

Regards,
Janne

[edavid]

Yeah, soon as the idea occurred I realised that the CRT would be the limiting factor, so 1Ghz CRT is as high as was ever produced? 

The CRTs in the Tek 7904A and Tek 7104 were good to at least 2GHz, but were limited by the electronics.  I'm sure a new design could be faster.

There is a French company called Greenfield Technology which sells a 7GHz CRT-based transient digitizer.  That CRT is easier to build than a direct-view tube, since the vertical deflection is smaller, and it uses a silicon target instead of phosphor.

http://www.greenfieldtechnology.com/-Data-aquisition-system-.html