EEVblog #311 - Jim Williams Pulser Followup

[EEVblog]

Dave.

[electrode]

Is he gonna be okay, doctor?

[KTP]

I am not going to hook up 7V to my 50ohm input when I get my MSOX3024A...I bet it has similar limits to that 13GHz scope (5V max)

I am thinking when you mentioned in the other video that it can handle 300V you probably meant with the 1M input impedance.

[olsenn]

Dave; where the f&#@ did you get that 13GHz DSO??? I'm assuming that isn't your personal device?

As for the 5V max input, I have a similar constraint on my DSA815 spectrum analyzer; I just attached a 30db attenuator to the input and leave it there. Honestly, I can't tell why low-Z is even required as the noise floor isn't greatly increased when I add a large resistor (especially when I turn the preamp on). I realize my device only goes to 1.5GHz and not 13GHz, but even still, most 50-ohm scopes are not much higher bandwidth than 500MHz and a DSO could easily read 1GHz+ signals with a 100k+ input impedance.

[StubbornGreek]

Crapped my pants at 30 seconds... 

The 'stand-in' scope was good enough to fuel round two, lol.

Very informative, thank you - as always.

[Rufus]

I am not going to hook up 7V to my 50ohm input when I get my MSOX3024A...I bet it has similar limits to that 13GHz scope (5V max)

The limit is printed on the front of the scope 5v rms which is 0.5W input power.

higher bandwidth than 500MHz and a DSO could easily read 1GHz+ signals with a 100k+ input impedance.

The Agilent 2GHz active probe Dave opened in the last mailbag vid costs more than a DSA-815-TG and has an input capacitance of 1pF. 1pF at 1GHz has an impedance of 159 ohms. The capacitance of just a BNC plug and socket is around 3pF. It is practically impossible for anything to present a 100k input impedance at 1GHz.

Oh and Dave - lol at the size of that attenuator.

[alm]

Honestly, I can't tell why low-Z is even required as the noise floor isn't greatly increased when I add a large resistor (especially when I turn the preamp on). I realize my device only goes to 1.5GHz and not 13GHz, but even still, most 50-ohm scopes are not much higher bandwidth than 500MHz and a DSO could easily read 1GHz+ signals with a 100k+ input impedance.

Hi-Z scopes have something like 8 pF of input capacitance. At 1 GHz, this has an impedance of about 20 ohm. A 10x probe will have an impedance of at best 200 ohm (usually worse since it also adds its own capacitance). I believe Tektronix is currently the only one selling a 1 GHz Hi-Z probe. What's the point of a 100 kohm resistor with effectively a 20 ohm resistor in parallel? Not to mention that 100 kohm transmission lines are somewhat impractical. If you want a higher input impedance, use an active probe (which can have a much lower input capacitance because the amplifier is right at the probe tip) into a 50 ohm input.

[EEVblog]

I believe Tektronix is currently the only one selling a 1 GHz Hi-Z probe.

I showed an Agilent 2GHz 1M/1pF probe in the latest mailbag video.

Dave.

[alm]

Let me rephrase that: I believe that Tektronix is currently the only one selling a 1 GHz probe that works with a hi-Z input. Active probes beyond 500 MHz have existed since before you were born (I think they can have bandwidths of tens of GHz now), but passive hi-Z probes were limited to 500 MHz until a few years ago. The fast active probes are usually limited to 50 ohm inputs, hence the need for low impedance inputs on faster scopes.

[KTP]

The limit is printed on the front of the scope 5v rms which is 0.5W input power.

What does the input circuitry of the Agilent 3000x series look like?  Is there some sort of transorb when it is switched to 50ohm, or is the danger just exceeding the power dissipation of the termination?  2pf charged to 7V discharged over 4ns isn't going to cause any thermal issues, but I don't know what other circuitry might be unprotected at the 50 ohm setting on the scope...

[vk6zgo]

I am not going to hook up 7V to my 50ohm input when I get my MSOX3024A...I bet it has similar limits to that 13GHz scope (5V max)

I am thinking when you mentioned in the other video that it can handle 300V you probably meant with the 1M input impedance.

I believe you are correct,as 300V RMS across 50 Ohms is 1.8kW!!

[KTP]

I am not going to hook up 7V to my 50ohm input when I get my MSOX3024A...I bet it has similar limits to that 13GHz scope (5V max)

I am thinking when you mentioned in the other video that it can handle 300V you probably meant with the 1M input impedance.

I believe you are correct,as 300V RMS across 50 Ohms is 1.8kW!!

I did a bit of a double take when Dave said "we are used to scopes that can handle 300V"

Still, I wonder if there really is any harm putting a low duty cycle 7V, 10V or even 90V pulse into the 50 ohm input if it is coming from a 2pF cap...lets see...using energy stored = (CV^2)/2 that would be about 8 nanojoules....probably not going to hurt a 50ohm resistor...

[Psi]

Still, I wonder if there really is any harm putting a low duty cycle 7V, 10V or even 90V pulse into the 50 ohm input if it is coming from a 2pF cap...lets see...using energy stored = (CV^2)/2 that would be about 8 nanojoules....probably not going to hurt a 50ohm resistor...

Yeah, 10V with such a tiny amount of energy is not going to do anything bad.

The real issue is... are you going to take the risk with a $140,000 scope that's on loan   

[KTP]

Yeah, 10V with such a tiny amount of energy is not going to do anything bad.

The real issue is... are you going to take the risk with a $140,000 scope that's on loan   

You ask this about a guy who tosses a 500mhz scope on the floor and drop tests a $30,000 MDO? 

I am quite suprised he didn't end the episode with a car driving over the 13 ghz unit.

[Psi]

You ask this about a guy who tosses a 500mhz scope on the floor and drop tests a $30,000 MDO? 

The fall was obviously staged, probably to counter all the youtube hate comments about the techtronic vs car video... Time to drop an Agilent to make things even

Unless of course if that was the intention and it missed the mattress or the person off-screen failed to catch it, which would be sad.

[vk6zgo]

I am not going to hook up 7V to my 50ohm input when I get my MSOX3024A...I bet it has similar limits to that 13GHz scope (5V max)

I am thinking when you mentioned in the other video that it can handle 300V you probably meant with the 1M input impedance.

I believe you are correct,as 300V RMS across 50 Ohms is 1.8kW!!

I did a bit of a double take when Dave said "we are used to scopes that can handle 300V"

Still, I wonder if there really is any harm putting a low duty cycle 7V, 10V or even 90V pulse into the 50 ohm input if it is coming from a 2pF cap...lets see...using energy stored = (CV^2)/2 that would be about 8 nanojoules....probably not going to hurt a 50ohm resistor...

Yeah,good point--you might take the chance if it was your own 'scope!
Of course,the rating is for worst case conditions,like a 5v RMS sinewave,or even DC.

With external terminations,as a lot of us use, if we accidentally put 300V across it,we'd just say,
"Bummer!.I've cooked the termination--where's the other one!",& dig through the junkbox till you find it.

[EEVblog]

Yeah, 10V with such a tiny amount of energy is not going to do anything bad.

Pointless even trying when it's only got 1V/div max.

Dave.

[tnt]

The 3000-X has a 5V RMS max when on 50 ohm (it written on the front panel) but I suspect this is just the power inside the termination ... I doubt the energy of the pulse would do any real damage.

[SeanB]

Might also be that on terminated position it bypasses some input protection, and excursions above this would damage an input amplifier stage.

[tnt]

Wouldn't they specify a peak voltage in that case as well ? My SA has both a RMS max power and a peak DC spec.

From the little I know bypassing input protection would be useful to get better front end bandwith but here you have the same specs for 50R and 1M. What would be the benefit of disabling them ? Better matching across all the bandwidth maybe ?

[SeanB]

I would suggest using a SMA connector to connect the coax, you are seeing the end termination in the beginning, and the connector will give a constant impedance for the initial connection.

[hans]

What I think more is a possibility of the low input limits is just protecting the silicon (well if they use it on this 13GHz model). I think it's incredibly hard to protect GHz signals. An protection diode may be ruining the signal performance. Especially at 13GHz , just an extra trace to a diode can ruin the input circuitry.

A class mate of mine got a HP network analyzer for free from college. The RF output was broken and the power supply failed.He fixed the power supply, some feedback opto's were dead. RF output was broken because some amplifier shorted out one of it rails, and the bond wires were gone. Someone probably connected an output signal of a network to the RF output. Bad luck, the RF module only costs a couple of thousand dollars. Ah well, he has a nice spectrum analyzer now though.
So, RF equipment are just very sensitive. Like, you also don't turn on a RF modem without an antenna right?

+/-5V sounds sufficient for differential signals, which I think this scope is designed for (measuring quality of transmission lines)
Whilst Dave got it, maybe he can do some signal testing with his Marconi function generator, and transmission lines? Just watch out for overshoot

[A Hellene]

Dave, can I have that cheep and useless 500 MHz item you trashed on the floor, in the beginning of the presentation (at approx. 0:00:30)?


-George

[Ketturi]

Well now we know how this little tester can perform i tough getting it and test my old Kenwood CS-5140 40Mhz analog higher-end(?) scope. Well it have some kind of sampling functionality from 40Mhz to 100Mhz range, but still i don't think it will be much better than cheap rigols. Does same practice as using DSO work with analog scope, at least i guess so.

But 13Ghz, wow, that is some naughty scope, I haven't even seen devices using so high frequencies. It would be really interesting see teardown, what magical stuff that box involves. But well, someday in future when we have flying DeLoreans and 13Thz scopes that equipment will be as ancient as WWII era scopes now.

[jahonen]

I have been reading Howard Johnsons book "High-speed signal propagation" and he mentions that skin effect induced loss affects the waveform so that it has relatively sharp edge but then slowly rising top. I wonder if those waveforms are affected by the skin-effect loss. Jim Williams used hardline coax in AN-94 (along with other edge-shaping components), which has much lower loss. That would be interesting to see. or even semi-rigid one.

Regards,
Janne

[RRobot]

With regards to the problem you had with the 2 nS vs 3 nS expected rise time on the Rigol, would this have to do with the 1GSa/ps limit on the Rigol introducing up to 1 nS of error? Was this with real or equivalent time sampling turned on?

[tinhead]

With regards to the problem you had with the 2 nS vs 3 nS expected rise time on the Rigol, would this have to do with the 1GSa/ps limit on the Rigol introducing up to 1 nS of error? Was this with real or equivalent time sampling turned on?

the error on Rigol E is probably coming from the DSO internal attenuator. While testing the bw with Marconi RF generator
the internal attenuator was off (200mv/DIV), but while testing the bw with pulse generator the attenuator was on (2V/DIV).

From my experience by default all DSO having this particular frontend (Rigol E, ATTEN, Siglent, Tekway/Hantek, etc.)
are calibrated to overshoot a bit while the internal attenuator is enabled, this helps out with the large signals response.

Due this overshoot the resulting bw (calculated back from rise time) seems to be "higher", which of course is not the real case.

[RRobot]

With regards to the problem you had with the 2 nS vs 3 nS expected rise time on the Rigol, would this have to do with the 1GSa/ps limit on the Rigol introducing up to 1 nS of error? Was this with real or equivalent time sampling turned on?

the error on Rigol E is probably coming from the DSO internal attenuator. While testing the bw with Marconi RF generator
the internal attenuator was off (200mv/DIV), but while testing the bw with pulse generator the attenuator was on (2V/DIV).

From my experience by default all DSO having this particular frontend (Rigol E, ATTEN, Siglent, Tekway/Hantek, etc.)
are calibrated to overshoot a bit while the internal attenuator is enabled, this helps out with the large signals response.

Due this overshoot the resulting bw (calculated back from rise time) seems to be "higher", which of course is not the real case.

I would not have thought of that. It would be interesting then to see the results with a 20dB attenuator and 200mV/dv settings.

[jpb]

With regards to the problem you had with the 2 nS vs 3 nS expected rise time on the Rigol, would this have to do with the 1GSa/ps limit on the Rigol introducing up to 1 nS of error? Was this with real or equivalent time sampling turned on?

The error shouldn't be as much as 1nsec in that for a rise time of 2.5 nsecs there should be either 2 or 3 samples on the slope but it will definitely increase the likely error. Another difference might be the interpolation - Agilent uses sin(x)/x which effectively gives 10 times as many points while the Rigol may just use linear.

On the circuit itself, I'm probably being naive, but to both reduce the magnitude of the pulse and increase the RC time constant to make it longer couldn't you just put say a 1k resistor in series with the 50 ohms? The pulse would then be less than 1V and the decay time constant would be 2nsecs or so. At the risk of damaging the transistor, the capacitor value could be made 5 or 10 pF to increase the time constant further and the pulse should not rise above 5V. Parasitic capacitance or inductance associated with the extra resistance might increase ringing but on the other hand the rest of the circuit would be further isolated from the 50 ohms by an extra 1k. It shouldn't affect the rise time of the pulse (I don't think).

[hobbs]

Nice scope.  Of course, my Tek 11801C sampling scope plus a few nice SD-24 plugins has almost twice the bandwidth, (20 GHz, 17 ps rise time) and I paid $1500 for it on eBay a month or two ago.  The faster plugins can go 50 GHz.  The SD-24 also has a 17-ps TDR pulse generator that I use for _everything_--even driving diode lasers.  (You bias the diode laser to just about threshold, and the 8 mA TDR pulse makes a nice fast optical edge--around 80 ps for garden-variety TO-can diode lasers.)

For repetitive signals, sampling scopes are the bomb. 

Cheers

Phil Hobbs

[rr100]

I just hope you have insurance for all that nice gear stashed around ... wherever you stash it.

[jpelczar]

Why does the attenuator have series of parallel discs ? I've seen some of the Nokia videos, where they measure something in 4G BTS, and they also used the attenuators withc discs.

[Rufus]

Why does the attenuator have series of parallel discs ? I've seen some of the Nokia videos, where they measure something in 4G BTS, and they also used the attenuators withc discs.

It is a high power attenuator and the 'discs' are heat sink fins.

[phenol]

   I know that this topic is antique, but still....i find that pulser with the associated boost circuit rather intriguing so i thought i should give it a try.

I made the booster with the well known and ubiquitous MC34063 chip and a super tiny handwound flyback transformer. It delivers regulated 95V and operates anywhere from 3 to 12-15V. The entire setup draws around 7mA tops.

The pulser itself is a BFR93A transistor retrofitted on some microwave Agilent demo board so that i could use the 50ohm microstrip lines already present there. The storage cap is 0805 2.7pF.

The two images show the spectrum picked up with a special E near field probe directly on the output microstrip line, the end of which is terminated with 2x 100ohm resistors. The pronounced notches in one of the scans are caused by the 30-40cm coax stub soldered across the storage 2.7p cap. the other scan is the plain circuit itself, no coax.

bye