Keysight MSOX-6004A Test to Impress Win

[rx8pilot]

Many of you have seen the various threads on the Keysight Technologies 'Test to Impress' contest where the bold and beautiful MSOX-6004A 6Ghz scope was the top prize. In the end, the the Keysight team decided to give away not 1, but 3 of these because the contest was so close. I am the very lucky recipient of this instrument and look forward to using its capabilities to push myself personally and my business to design more advanced products.

Almost the day is showed up, I have been buried in manufacturing a rush on our main product and unable to do much with the new scope. That rush is starting to slow, and I wanted to have a place to discuss some of the details of what this scope can do. There is no question that a number of the features are beyond what I have used before and therefore hoping that this can be a good place to look at the details and better understand how to use it to gain insight into a design challenge.

One hurdle that will take a while to get over is that I need a number of probes to do what I plan to do with this. At the moment, I only have the 10:1 500Mhz passive probes that are kitted with it. I am looking at the various active high-speed voltage probes as well as current probes. The total on the probe wish list is about $15k which is not as bad as it sounds since I was pretty close to spending that on a new scope package anyway. Now, that money can go toward accessorizing this one.

In the meantime, I am trying to gather the video from the delivery (which was a TON of fun!) and fiddling with the general operation. I have connected it to a real PCB once so far. As I dig in, questions will come up and hopefully a discussion here can help get to the next level. So much to learn. So much fun!

Fun note: This is my first ethernet enabled piece of test equipment. Goodbye GPIB!

[timb]

Man, that is a gorgeous piece of kit! Looking forward to seeing a video of it in action.

Also, thank you once again for passing your old scope on! I can't properly express how grateful I am for that. I'll be putting it to good use and hopefully it'll serve me as well as it did you!

So, what kind of active probes are you looking at for the new scope? What sort of bandwidth?

[rx8pilot]

Man, that is a gorgeous piece of kit! Looking forward to seeing a video of it in action.

Also, thank you once again for passing your old scope on! I can't properly express how grateful I am for that. I'll be putting it to good use and hopefully it'll serve me as well as it did you!

So, what kind of active probes are you looking at for the new scope? What sort of bandwidth?

Thank you timb - I need to go buy some packing material to ship your scope. The boxes I have are too small and I was not able to leave my desk all last week.

For probes - I need a variety to do what I want to do.

SMPS design/validation I need a pair of DC current probes, differential amplifier or probes for sense resistor current measurements, injection transformer for loop response measurements, and likely the power-rail probe.
For high-speed digital I need high-speed 6Ghz active probes to look at timing and signal integrity on signals that will probably be close to the BW limit of this scope. I suspect there will be a measurement technique learning curve to go along with my circuit design and PCB layout learning curves in the high-speed realm.

First task is to figure out how the scope deals with memory. I did a few tests and could not figure out how to capture more data than what fits on the screen. It is important that I understand the details of how the sample rate and memory are handled.

Basic functions seem to be quite obvious and intuitive and I love having the choice of buttons, touch screen, and mouse to navigate. I was quickly able to get to the point where I was functionally with my old Tek scope in terms of basic setup, triggering, horiz/vert controls, etc. The fancier features like reference wave forms, trigger on data packets, dual channel arbitrary waveform generators that can take a displayed waveform and 'play' it back. Awesome.

Before I can even think of spending $15k on probes and other accessories - I need to master the fundamentals. Maybe some of the other winners of the 3000's, 4000's and other Keysight owners can chime in. Maybe I should post video questions too.

[Berni]

Such a beautiful piece of kit that thing. Makes my old MSO6000 look like a toy.

Probes for getting the most of of such a scope are going to be expensive alright. I have a 4GHz MSO9000 that i use for high speed digital. My favorite probes for it are the HP 1152A active 2.5GHz probes. They can be had for 150 bucks on ebay and feature very low tip capacitance so that you don't mess up your circuit under test. In most cases 2.5GHz is plenty and the probes are pretty solid so should survive a lot of every day use as long as you don't drop one on the floor as the ceramic PCB inside is fragile. They have a common mode reach up to +/- 10V and are specified to survive 30V. Only problem is that these are infiniium branded probes while your scope might only support infinimax type probes. But there is a trick of covering the ID pin with tape to prevent the scope from detecting it (or if you change the resistor inside it will detect it as a diferent probe and work fine) and i got them to work on my 6000, just the voltage range of the probe is not to spec (The supply voltage is lower on non infiniium scopes)

As for using your scope to look at USB 3.0, HDMI etc then you will need a 1134A. These work on all modern scopes(Tested one on mine 6000). Those are significantly more expensive tho but you don't need more than one.

With about 2 grand and some time on ebay you can get all the basic probes to put that massive bandwidth to use.

Have fun with the new toy.

[HighVoltage]

Great that you have this beauty in your lab now.

Once in a while there are really good deals on probes on ebay.
But you have to have patience for a long time.
Congratulations again.

[rx8pilot]

Patience has served me well so far....a requirement for eBay shopping. The next few projects I need to get done, will not need the high speed - so I am first looking for the current probes to get some power electronics through the process. The scope includes the the POWER analysis software option that I very much look forward to using.

One of the coolest option to be integrated is the FRA (frequency response analyzer) that uses the built in sig-gen to create a poor-mans FRA. While it may not be as sensitive and accurate as a Venable, it is FAR better than guessing the performance of the loop in SMPS designs. The POWER option also includes more basic functions that seem like they will speed up the process. I was previously using differential probes and sense resistors for current monitoring, but had to manually calculate everything. The Keysight scope with the POWER option automates the process. Perfect.

The only thing I worry about in working with SMPS designs is that it is more likely to make a probing mistake that can damage the scope. When I had a far lower cost scope - I was less nervous. It is quite nice to have a scope that can service needs in power and high-speed at the same time - just need to be bonus careful.

My favorite probes for it are the HP 1152A active 2.5GHz probes.

Not a bad suggestion. The cost is low enough that they can be the 'everyday' digital probes, where the juicier 6Ghz differentials would be more cautiously deployed. As for current probes - I am generally needing modest current <15A max and commonly need (want) to see relatively fine resolution of mA with lower currents of only a few amps. Using tiny 5mOhm sense resistors and differential amplifiers was ok for that. The ultimate accuracy was not likely all that good but generally I was looking for relative measurements.

I feel like I should have both clamp-on
http://www.keysight.com/en/pd-2202912-pn-1147B/50-mhz-15a-ac-dc-current-probe?cc=US&lc=eng

 and the resistor based N2820A
http://www.keysight.com/en/pd-2229629-pn-N2820A/3-mhz-50ua-high-sensitivity-ac-dc-current-probe-2-ch?cc=US&lc=eng

[nctnico]

IMHO a sense resistor and amplifier is a very good way to measure current accurately (ofcourse depending on the sense resistor and amplifier accuracy but these are easy to control). Clamp-on current probes use hall-effect chips and current transformers. Generally speaking they have large offsets and lots of noise.

[rx8pilot]

IMHO a sense resistor and amplifier is a very good way to measure current accurately (ofcourse depending on the sense resistor and amplifier accuracy but these are easy to control). Clamp-on current probes use hall-effect chips and current transformers. Generally speaking they have large offsets and lots of noise.

Maybe I should consider a pair of the Preamble/LeCroy 1855 differential pre-amps. It seems the specs are rather awesome - much better than I had before. There are more uses other than current sensing too - bonus for versatility.

[H.O]

Maybe I should consider a pair of the Preamble/LeCroy 1855 differential pre-amps.

Isn't that new baby of yours four channels? I'm sure two of them will feel left out if you get diff amps for only two of them :-)

How much are you looking to spend anyway? Same amount as that scope would've cost you had you not got it for free? Congrats BTW!

[rx8pilot]

Maybe I should consider a pair of the Preamble/LeCroy 1855 differential pre-amps.

Isn't that new baby of yours four channels? I'm sure two of them will feel left out if you get diff amps for only two of them :-)

How much are you looking to spend anyway? Same amount as that scope would've cost you had you not got it for free? Congrats BTW!

Diff amp on two channels and two voltage probes on the other. Full house 

I was anticipating around $15k if I get a combination of new and used-but-nearly-new. That would cover the power related and high-speed related needs that I can think of now. Obviously, if I get lucky and find deals that will be great and I do have some time to build up a probe package for this scope. It is a LOT cheaper when you are not in a huge hurry.

[rx8pilot]

I connected the scope to one of my cameras last night just to see some high-speed action (sort of).

1.485 Gbps HD-SDI which is a serial signal on a 75 Ohm coax cable. I terminated it at 50 Ohms, but it still shows some fast edges that I could not see before. The rise time is about 120ps or so and the signals I am most interested in are about 60-75ps. I was able to do some jitter analysis and build a real time eye too. This is getting fun.

[Howardlong]

Here's a link to the 1152a mod. https://www.eevblog.com/forum/testgear/agilent-mso-500060007000-anyone-hacked-these-scopes/msg868073/#msg868073

It may be that your scope supports this probe directly without any fiddling about, but just be aware that the significant DC offset I encountered on my x3054a, and there isn't a means to adjust it from the UI, so it means I don't generally use it on that scope. It is, however, quite a bargain for a FET active probe. There may be a way to adjust the offset by adding a trimpot, maybe on pin 3 of the autoprobe interface?

[rx8pilot]

Here's a link to the 1152a mod. https://www.eevblog.com/forum/testgear/agilent-mso-500060007000-anyone-hacked-these-scopes/msg868073/#msg868073

It may be that your scope supports this probe directly without any fiddling about, but just be aware that the significant DC offset I encountered on my x3054a, and there isn't a means to adjust it from the UI, so it means I don't generally use it on that scope. It is, however, quite a bargain for a FET active probe. There may be a way to adjust the offset by adding a trimpot, maybe on pin 3 of the autoprobe interface?

That is juicy news and a fast/easy/low-cost path to get 2.5Ghz single ended probes. GREAT info.

[Berni]

Who knows if you are lucky they might work unmodified on that scope. Only one way to find out.

Getting under 100ps rise time signals is not easy by the way, Perhaps looking at some USB 3.0, but you might need a proper probing solution for that.

[rx8pilot]

Getting under 100ps rise time signals is not easy by the way, Perhaps looking at some USB 3.0, but you might need a proper probing solution for that.

Yes, agreed. <100ps is nuts. I expect that the N2752A 6Ghz is the best option to deal with those signals properly. These signals are generally only seen in broadcast TV and film production and post production. My new designs will need to generate and receive these signals to the appropriate SMPTE (Society of Motion Picture Television Engineers) specification. Line drivers and equalizers are generally off the shelf silicon and that places my primary challenge on PCB layout, connectors, power etc. I need to be able to quantify delay and jitter as well as seek out any infrequent glitches, runts, or other error generating anomalies. This scope is perfect for that - but only with the right probes for sure.

I suspect that placing dedicated high-speed probe points on the PCB would be very wise. Hoping the 'technique' learning curve is modest trying to look at signal integrity issues at close to the limits of the BW of this scope.

For those that missed the 'Arrival' video of this scope......
https://youtu.be/_GIldByHYPk

[nctnico]

IMHO a TDR and impedance analyser (or spectrum analyser + generator + directional coupler) are much better tools for validating transmissions lines than a scope when using off-the-shelve silicon. A TDR allows you to see any impedance changes versus time so gives a good view on what the signal path looks like. An impedance analyser shows the impedance versus frequency which gives some insight into signal attenuation over frequency (see also SWR https://en.wikipedia.org/wiki/Standing_wave_ratio ).

[tautech]

Patience has served me well so far....a requirement for eBay shopping. The next few projects I need to get done, will not need the high speed - so I am first looking for the current probes to get some power electronics through the process. The scope includes the the POWER analysis software option that I very much look forward to using.

One of the coolest option to be integrated is the FRA (frequency response analyzer) that uses the built in sig-gen to create a poor-mans FRA. While it may not be as sensitive and accurate as a Venable, it is FAR better than guessing the performance of the loop in SMPS designs. The POWER option also includes more basic functions that seem like they will speed up the process. I was previously using differential probes and sense resistors for current monitoring, but had to manually calculate everything. The Keysight scope with the POWER option automates the process. Perfect.

Be sure to Deskew any propagation delays out between current and voltage probes for PA computations to be accurate.

[rx8pilot]

IMHO a TDR and impedance analyser (or spectrum analyser + generator + directional coupler) are much better tools for validating transmissions lines than a scope when using off-the-shelve silicon. A TDR allows you to see any impedance changes versus time so gives a good view on what the signal path looks like. An impedance analyser shows the impedance versus frequency which gives some insight into signal attenuation over frequency (see also SWR https://en.wikipedia.org/wiki/Standing_wave_ratio ).

While those are good to consider for those tasks, I don't have any of them yet. For the transmission line signals, to some extent, I will have to go with the reference designs and use the scope to validate the signals as best as possible. Since the parameters of the transmission line and the signals on it are etched in stone - should I really be worried about  analysis at that level? The connectors, layouts, external components have long since been chosen. Is that a logical fallacy?

Most of the signal measuring will happen differentially on the PCB and need to see noise, phase, and other signal anomalies as well as measure timing from the FPGA's driving the signals - things that scopes are good at.

Be sure to Deskew any propagation delays out between current and voltage probes for PA computations to be accurate.

Johnnie Hancock from Keysight gave me a lesson on this personally while doing a demo on the POWER option. It's an easy detail to overlook so the scope actually reminds you during the setup of the processes that use the current probes. Great point - skipping that step could have me chasing my tail indefinitely.

[rx8pilot]

IMHO a TDR and impedance analyser (or spectrum analyser + generator + directional coupler) are much better tools for validating transmissions lines.......

You have already got me looking at new gear with all the money I saved winning the scope 

[Brumby]

I am still in awe of the screen real estate ... I mean, that's acreage!

I will now step back - and put my voyeur hat on........

[rx8pilot]

The display is indeed enormous and awesome.

Sent from my horrible mobile....

[rx8pilot]

Trying to understand the limitations of the FFT. I have never owned an SA, so my ability to compare is near zero. First look seems like it is slow and has a high noise floor - but keep in mind that this is a new place for me that I have very little understanding and not even sure where to look first.

The options seem minimal. Does anyone have tips/hints/white papers/links/etc for using FFT? I had generally ignored FFT since I did not have the capability, but with a 6Ghz-20GS/s front end I was hoping it would allow me to get through some pre-compliance and very basic RF measurements (with external amplifiers). Curious how much mileage I can get out of the FFT.

How to choose the window? What noise floor should I expect in a best-case scenario? The no-input noise floor seems about -97dBm.

EDIT:
As usual, a great video from W2aew
https://youtu.be/oRf-IpG6XAw

[tautech]

There's a good paper linked by rf-loop in this post:
https://www.eevblog.com/forum/testgear/siglent-ssa3000x-spectrum-analyzers/msg965873/#msg965873

[tautech]

And another linked by rf-loop somewhere that I could not find:
https://cdn.rohde-schwarz.com/pws/dl_downloads/dl_application/application_notes/1ef77/1EF77_3e_Real-time_Spectrum_Analysis.pdf

[Berni]

You can get a bit better FFT with averaging on the waveform and playing with the settings just right. The 8 bit ADC in there is the limiting factor with its quantization noise. But with all those GS/s the high res mode is pretty useful. I only ever saw this scope at a Keysight stand on a trade show(and the guy there quickly tried to sell me on one) so I don't know how high res works on it, but my 9000 has a useful one where it down samples the signal by a fixed ratio. For example it turns 20GS/s in to 2,5GS/s with 500MHz of bandwidth and what they claim 12bit resolution. It does make the signal nice and clean without the weird artifacting that the old scopes do in high res. Since a lot of times I don't need more than 500MHz I usually run it in this mode.

Spectrum analyzers only have better noise floors when using low RBW settings. I'm still getting the hang of those since I bought my first SA just a few days ago (never did much RF work so have t had the need for one)