Something amazing! Keysight UXR 110GHz, 256GS/s, 10-bit Scope Teardown & Exp.

[Hugoneus]

To answer a few more questions:

1) Waveguide connectors would not make good scope inputs since going down to DC is required. If Keysight were to push the bandwidth even higher, they would have to go to sub-1mm connectors such as what is used by Anritsu.

2) The scope was delivered by FedEx and picked up by FedEx. Nothing unusual or out of the ordinary. The box is huge! Total weight was about 70kg (160 lb).

3) I don't know the internal deals between Micron and Keysight. But I am sure that this is something they have worked out internally.

4) A small error in the video I made is that the smallest hardware vertical division is 7.5mV/Div for InP input and 4mV/Div for SiGe input. But of course the noise measurements are correct and valid.

[srce]

It sounds like Micron is discontinuing the HMC memory in use by this scope - I wonder how it will impact what Keysight is doing. If HMC is a key enabler for their tech and baked into their digital ASIC (which I suspect they want to use for a long time to come given what it would have cost to develop) then have they got a major issue?

(see https://www.micron.com/support/faqs/products/hybrid-memory-cube and the note about last-time-buy etc)

That's a shame. It says they now have a HBM development program, but HMC looks a lot easier to integrate. 

[tv84]

A key component discontinued in a brand new million-dollar equipment...     

[Fungus]

It sounds like Micron is discontinuing the HMC memory in use by this scope - I wonder how it will impact what Keysight is doing. If HMC is a key enabler for their tech and baked into their digital ASIC (which I suspect they want to use for a long time to come given what it would have cost to develop) then have they got a major issue?

They can probably stockpile some. It's not like they'll need millions of chips.

[drussell]

A key component discontinued in a brand new million-dollar equipment... 

I'm sure they worked closely with Micron on the memory subsystem design.  They will have either already secured a lifetime-buy of the necessary components or will have an upgrade path in the works, I would expect.

[Fungus]

2) The scope was delivered by FedEx and picked up by FedEx. Nothing unusual or out of the ordinary.

[drussell]

The question should be, "How much is the insurance on shipping a $1.3 million item via FedEx?" 

[TK]

80% of the price must be R&D, Marketing, commissions, Fixed keysight costs, etc. etc. etc.

[Keysight DanielBogdanoff]

True story:

Last week we had a presentation held by Keysight (for the Device Current Waveform Analyzer).  I talked to the sales man about the UXR and he said there is no opportunity to
have a presentation outside of the Keysight labs with this oscilloscope (except you buy one).   

@hscade, I know that there will be demo scopes available, but as of right now (today) there aren't. It shouldn't be hard to get a demo, especially for one of the lower bandwidths once those are shipping.

[srce]

I see the software has had a phase noise application added. Will this work with the S series or is it just for the UXR? Can't see anything on the web site about it yet.

Nice to see SDA and UDF are now included as standard. 


I see the UXR still has a 1024x768 res display though  You quickly run out of real estate with all the different types of analysis you can do.

[maxwell3e10]

100 GHz - $1M
10 GHz - $100k
1 GHz - $10k
100 MHz -$1k
10 MHz -$100 (hand-held scope)
1 MHz -$10 (chip level)

1 cent per 1 kHz of bandwidth across 6 decades.

[TheUnnamedNewbie]

2) The scope was delivered by FedEx and picked up by FedEx. Nothing unusual or out of the ordinary. The box is huge! Total weight was about 70kg (160 lb).

It is at times like this that I wonder if he/she was aware that that box they were dropping off was probably worth more than they (or most other people, really) will ever see in their life. Reminds me of that time I was moving VNA extenders, and someone casually told me "If you drop that, it is pretty much the same as driving a brand new Porsche into a concrete wall"

Regarding the waveguide-part: I do find it hard to imagine we will see scopes with higher bandwidth than 110 GHz with 'simple' front-panel connectors. The loss of the cable will just be too much for longer runs. Perhaps a system similar to the millimeter-wave VNA modules will be the future, where the first bit of processing (attenuation, amplifier, and either S&H or mixers) happens in a 'smaller' external box right at the DUT, and a number of parallel channels will then connect that external box the scope, where the main A/D and DSP takes place. Of course, timing will be a mess in that case, but I am sure something can be engineered for that (some kind of sync system where the module will send a pulse over each channel to deskew? ).

[Mr. Scram]

80% of the price must be R&D, Marketing, commissions, Fixed keysight costs, etc. etc. etc.

Then again, this kind of research can't be cheap and there's not exactly a huge market for 110 GHz oscilloscopes.

[mikeselectricstuff]

1) Waveguide connectors would not make good scope inputs since going down to DC is required.

Do scopes like this actually need to go down to DC?

[TheUnnamedNewbie]

1) Waveguide connectors would not make good scope inputs since going down to DC is required.

Do scopes like this actually need to go down to DC?

One of the main uses of these scopes would be looking at digital signals, which depending on the coding might not have a strict DC component, but they will still go down to very low frequencies (in wave-guide terms). A waveguide around 110 GHz would be something like WR10, which only covers from 75-110 GHz. Until we find a way to get wider band waveguides that work well in production (eg, dielectric waveguides), we will have to stick to coax.

[Berni]

I see the software has had a phase noise application added. Will this work with the S series or is it just for the UXR? Can't see anything on the web site about it yet.

Nice to see SDA and UDF are now included as standard. 


I see the UXR still has a 1024x768 res display though  You quickly run out of real estate with all the different types of analysis you can do.

When you use Windows on a touchscreen its often a good thing to have pretty low resolution so you can hit the small buttons.

If you need more room you can always connect an additional DVI monitor and run it at any resolution you want (Tho maybe not 4K 60fps since the embeded PCs tend to be a bit old so it might not support the latest DVI spec). So far i have not seen multi-window support in the infinitum software, but it will run on any resolution you want. The VSA software might have some multiwindow support tho (The thing used for that QAM analysts in the video)

[drussell]

When you use Windows on a touchscreen its often a good thing to have pretty low resolution so you can hit the small buttons.

What would screen resolution have to do with being able to hit buttons?

If you doubled both the H and V resolution of your screen, for example, you would also double the H and V size of your buttons, so they would still be the same size, they would just have 4x as many pixels in them.

[Berni]

When you use Windows on a touchscreen its often a good thing to have pretty low resolution so you can hit the small buttons.

What would screen resolution have to do with being able to hit buttons?

If you doubled both the H and V resolution of your screen, for example, you would also double the H and V size of your buttons, so they would still be the same size, they would just have 4x as many pixels in them.

On these scopes you don't always just use the scope software in fullscreen and nothing else. Sometimes you also run just regular non touchscreen designed windows programs on it and the UI scaling feature in Win 7 are not always perfect.

For example when doing modulation analysis you tend to run this regular oldschool windows looking app on the scope:
https://www.keysight.com/zz/en/software/application-sw/89600-vsa-software.html

EDIT: Tho maybe they started going to Windows 10 now. The UI scaling is known to be much improved there.

[srce]

My S series has a Display Port out, so you can attach to a monitor at 2560x1440. And yes, the Infiniium s/w has multiple window support, so you can have different windows on each display  . The trouble with an external monitor, is that it's not quite in the position you want it and there's no touch screen support.

Yep, you do run other s/w on it, which is why a higher res is useful, so you can see multiple things at the same time.

They do now ship with Win10. But with regards to scaling, even Full HD, which wouldn't need scaling, would be a welcome improvement on 1024x768.

[David Hess]

If the last 25 years are anything to go by, 25 years from now this will be an entry level spec scope.

Entry level oscilloscopes have not changed much in bandwidth having gone from perhaps 50 to 100 MHz.  40 years ago the rough equivalent to this Keysight would have been some variation of a 14 GHz sampling oscilloscope which is still way beyond entry level and even most high end models.

It sounds like Micron is discontinuing the HMC memory in use by this scope - I wonder how it will impact what Keysight is doing. If HMC is a key enabler for their tech and baked into their digital ASIC (which I suspect they want to use for a long time to come given what it would have cost to develop) then have they got a major issue?

They could just do a lifetime buy.  The HMC memory cost is not significant compared to the rest.

[dr.diesel]

480v 3-phase?   

All joking aside, very impressive machine. 

[BrianHG]

When you use Windows on a touchscreen its often a good thing to have pretty low resolution so you can hit the small buttons.

What would screen resolution have to do with being able to hit buttons?

If you doubled both the H and V resolution of your screen, for example, you would also double the H and V size of your buttons, so they would still be the same size, they would just have 4x as many pixels in them.

On these scopes you don't always just use the scope software in fullscreen and nothing else. Sometimes you also run just regular non touchscreen designed windows programs on it and the UI scaling feature in Win 7 are not always perfect.

For example when doing modulation analysis you tend to run this regular oldschool windows looking app on the scope:
https://www.keysight.com/zz/en/software/application-sw/89600-vsa-software.html

EDIT: Tho maybe they started going to Windows 10 now. The UI scaling is known to be much improved there.

  OMG, how is this even a thing.  It's Keysight's own source code.  It they want a higher resolution, even if they were using Windows95 for their OS, in their scope code, they would change the font and button size and ignore Window's scaling BS, or, they have already programmed their software correctly so that even WindowsXP would scale their on screen elements correctly.  Why are we questioning this trivial thing on a 1.6m$ scope?
They actually know what monitor size and res is mounted in the scopes...  It's not like a separate purchased program on a random consumer home computer with a random setup.

[David Hess]

OMG, how is this even a thing.  It's Keysight's own source code.  It they want a higher resolution, even if they were using Windows95 for their OS, in their scope code, they would change the font and button size and ignore Window's scaling BS, or, they have already programmed their software correctly so that even WindowsXP would scale their on screen elements correctly.  Why are we questioning this trivial thing on a 1.6m$ scope?

They actually know what monitor size and res is mounted in the scopes...  It's not like a separate purchased program on a random consumer home computer with a random setup.

While I doubt it is a problem here, often the screen resolution of the graticule is made to be a fixed integer ratio of the digitizer resolution to prevent aliasing or the extra processing necessary to correct aliasing.  So for instance 200 counts from 8-bits might translate into 8 vertical divisions of exactly 25 pixels each and changing screen resolutions would produce a more smeared display.  Preserving this would require specific programming for different screen resolutions.

[egonotto]

Hello,

is in the datasheet © Keysight Technologies, 2018, Published in USA, September 19, 2018, 5992-3132EN an error?

Page 42 and page 43:
"
Offset range
Vertical sensitivity
1 mV/div to 55 mV/div
56 mV/div to 128 mV/div
129 mV/div to 278 mV/div
279 mV/div to 5000 mV/div" ??


On page 47: "Max time between triggers
> 100,000 years"
If i start to save money than I can perhaps buy an UXR1104A in 100000 years

Is the mean time to failure greater than 100000 years?

Best regards
egonotto

[mmagin]

480v 3-phase?   

According to the specs, It looks like the 2 channel version he looked at "only" requires 1370 VA and can run off 120V.  The 4 channel version only works off 200-240V

It blows my mind the extent to which the accuracy of something like this relies on both good RF design and some very clever DSP.  The whole time-interleaved sampling seems like it must rely heavily on the (DSP) software to calibrate out the subtle differences in all those signal paths.