NEW Keysight HD3

[tautech]

The SDS3000X HD is specced at 125uV RMS @1GHz, and it's not the best on the market.

It would also be quite embarrassing for others if the cheapest 1GHz scope was also the “best” in terms of noise on the market.
It's certainly not the worst either.

We wait and see..

What, any dimwit can use a screwdriver. 

[2N3055]

It would also be quite embarrassing for others if the cheapest 1GHz scope was also the “best” in terms of noise on the market.
It's certainly not the worst either.

We wait and see..

My SDS3000X HD has 130uV RMS @1GHz BW at 10 mV/div.. by the way..

The data points I've seen from the Siglent HD models overlaps identically with the Lecroy data above (unsurpisingly).

Good observation. Some do, some don't.

Scope's front end is designed according to input sensitivity/offset/amp stages/attenuator stages plan.. If you have similar plans numbers start to converge.
Being OEM for some LeCroys and long cooperation with them will definitely have influence to "flavour" of scopes Siglent makes.
But there are some Siglent scopes  that have front ends that are unique, having large offset ranges not many others have...

[EEVblog]

The SDS3000X HD is specced at 125uV RMS @1GHz, and it's not the best on the market.

the SDS3104X HD has less than 6 mV RMS noise at 1 GHz and 50 ohms and 1 V/div and 1 ms/div and 40 Mpts and 4 Gpts/s. Only LeCroy can beat that. With other time and memory settings, the SDS3104X HD can also be brought below 5 mV noise.

When I get it, I'll measure it. But according the specs, the noise is indeed half of the best competitor on the 2mV range. 1V is not specified in the info I have.
And in case anyone has missed mention of it, the HD3 has more resolution than all the competitors.

[2N3055]

The SDS3000X HD is specced at 125uV RMS @1GHz, and it's not the best on the market.

the SDS3104X HD has less than 6 mV RMS noise at 1 GHz and 50 ohms and 1 V/div and 1 ms/div and 40 Mpts and 4 Gpts/s. Only LeCroy can beat that. With other time and memory settings, the SDS3104X HD can also be brought below 5 mV noise.

When I get it, I'll measure it. But according the specs, the noise is indeed half of the best competitor on the 2mV range. 1V is not specified in the info I have.
And in case anyone has missed mention of it, the HD3 has more resolution than all the competitors.

All scopes will have different sweet spots..

14 bits sounds nice, ENOB is better...
So we are saying 50-66µV RMS at 1GHz BW at 2mV/div....

As I said we'll see in few weeks. I don't expect Keysight to do it incompetently. We'll see how far they could push it though.
 

[nctnico]

The SDS3000X HD is specced at 125uV RMS @1GHz, and it's not the best on the market.

the SDS3104X HD has less than 6 mV RMS noise at 1 GHz and 50 ohms and 1 V/div and 1 ms/div and 40 Mpts and 4 Gpts/s. Only LeCroy can beat that. With other time and memory settings, the SDS3104X HD can also be brought below 5 mV noise.

When I get it, I'll measure it. But according the specs, the noise is indeed half of the best competitor on the 2mV range. 1V is not specified in the info I have.
And in case anyone has missed mention of it, the HD3 has more resolution than all the competitors.

Unless it is some kind of high-res mode. Realistically, having more than 12 bits with hundreds of MHz of bandwidth requires an insanely low jitter / low phase noise sampling clock. Otherwise the jitter / phase noise by itself messes up the ENOB big time. So for actual measurements it makes sense to trade bandwidth for more bits / resolution.

[EEVblog]

Unless it is some kind of high-res mode. Realistically, having more than 12 bits with hundreds of MHz of bandwidth requires an insanely low jitter / low phase noise sampling clock. Otherwise the jitter / phase noise by itself messes up the ENOB big time. So for actual measurements it makes sense to trade bandwidth for more bits / resolution.

I'm sure Keysight are aware of this.

[mawyatt]

Could this be the First Commercial use of the Non-Uniform Sampling ADC championed by Dr Mike Chen at USC almost 2 decades ago

The NUS ADC certainly found uses it "other" places 

BTW KS was very interesting in this ADC, Dr Chen had a few visits from some very famous KS folks!!

Here's an image of an early rendition NUS ADC in TSMC 65nm CMOS

Later KS were developing interesting ADCs of their own, like this 2nd image

Best,

[EEVblog]

Here's an image of an early rendition NUS ADC in TSMC 65nm CMOS

Damn that's a gorgeous photo!

[tautech]

Could this be the First Commercial use of the Non-Uniform Sampling ADC championed by Dr Mike Chen at USC almost 2 decades ago

The NUS ADC certainly found uses it "other" places 

BTW KS was very interesting in this ADC, Dr Chen had a few visits from some very famous KS folks!!

Here's an image of an early rendition NUS ADC in TSMC 65nm CMOS

Later KS were developing interesting ADCs of their own, like this 2nd image

Best,

Not signed by yours truly ?

[mawyatt]

Thanks, we started developing Chip Imaging techniques and equipment back in 2000. Everything was custom developed from the lenses, setup, focus rails, even the lighting, software and so on!

Some folks wanted us to start company doing such, but we resisted. Becoming a "Chip Photographer" wasn't what we had in mind as our career wound down!!

We have a bunch of Specialized and Custom Chip images, some we still can't revel, but hopefully someday as we've recently been informed that some of our very old work may become public and put on display at a specialized museum., we shall see

Best,

[mawyatt]

Could this be the First Commercial use of the Non-Uniform Sampling ADC championed by Dr Mike Chen at USC almost 2 decades ago

The NUS ADC certainly found uses it "other" places 

BTW KS was very interesting in this ADC, Dr Chen had a few visits from some very famous KS folks!!

Here's an image of an early rendition NUS ADC in TSMC 65nm CMOS

Later KS were developing interesting ADCs of their own, like this 2nd image

Best,

Not signed by yours truly ?

 

Best,

[EEVblog]

Thanks, we started developing Chip Imaging techniques and equipment back in 2000. Everything was custom developed from the lenses, setup, focus rails, even the lighting, software and so on!
Some folks wanted us to start company doing such, but we resisted. Becoming a "Chip Photographer" wasn't what we had in mind as our career wound down!!
We have a bunch of Specialized and Custom Chip images, some we still can't revel, but hopefully someday as we've recently been informed that some of our very old work may become public and put on display at a specialized museum., we shall see

Tha reminds me, I have to arrange that Amp Hour podcast we were going to do!

[mawyatt]

Here's an image of an early rendition NUS ADC in TSMC 65nm CMOS

Damn that's a gorgeous photo!

Here's another Low Resolution Image of the NUS ADC you might like.

Best,

[tautech]

Here's another Low Resolution Image of the NUS ADC you might like.

Just distinguishable since I know what to look for, bottom right, correct ?

[mawyatt]

Here's another Low Resolution Image of the NUS ADC you might like.

Just distinguishable since I know what to look for, bottom right, correct ?

There's some text hidden under the wire bonds, if you look on the right middle there's more text. This is "MOSIS" which is chip fabrication aggregator, but the chip was actually fabricated at TSMC.

Best,

[mawyatt]

Hey Dave, if we're correct about the new Keysight HD3 using the NUS ADC, do we win the HD3 as a "prize" 

BTW interested folks should look into this NUS ADC. Simultaneously quantizing in Amplitude and Time opens up a completely new Signal Processing Path, like post Analog to Digital Conversion Anti-Aliasing Filtering

https://www.eevblog.com/forum/metrology/why-dont-we-see-more-tdc-based-high-resolution-adcs/msg5484382/#msg5484382

Best,

[EEVblog]

Hey Dave, if we're correct about the new Keysight HD3 using the NUS ADC, do we win the HD3 as a "prize" 

I asked them about that actually, and they said they'll ask R&D. So it sounds like like there is no marketing info boasting about it.

[nctnico]

Hey Dave, if we're correct about the new Keysight HD3 using the NUS ADC, do we win the HD3 as a "prize" 

I asked them about that actually, and they said they'll ask R&D. So it sounds like like there is no marketing info boasting about it.

I don't think a NUS ADC solves much problems except for bringing the anti-aliasing filter into the digital domain. I looked into it a bit and what a NUS ADC does is look at the input voltage continuously through the input comparator section and produce a value pair consisting of time and amplitude  when the comparator section senses an input voltage variation. However, quantizing time using a TDC (time to digital converter) for example, is bound to limits (as Mr Chen also notes in his video). Getting into the tens of fs time resolution with a TDC is hard. I know this because I had reason to ask this question to a specialist IC designer for a project I'm involved in. So in the end you still have what is an equivalent to aperture jitter which is in the same ballpark as a good conventional high speed ADC.

[jusaca]

But he also mentions some bold claimes of up to 30 dB noise advantage compared to classic ADCs, although I could not follow the reasoning while watching that IEEE presentation.

[pdenisowski]

I asked them about that actually, and they said they'll ask R&D. So it sounds like like there is no marketing info boasting about it.

In my experience, any ground-breaking / game-changing / unique features of a new product usually are very widely known within the company, especially less than a month before launch.  If someone has to ask R&D about a new product feature, it very likely either isn't there or isn't that important

[nctnico]

But he also mentions some bold claimes of up to 30 dB noise advantage compared to classic ADCs, although I could not follow the reasoning while watching that IEEE presentation.

The way I see this, NUS only has a benefit for relatively low frequency signals. If you can measure the time in single picoseconds while sampling a 10MHz signal, you have an effective samplerate of 1Gs/s without needing to process 1Gs/s. The extra effective samplerate gives you much more headroom in a digital anti-aliasing filter. The idea is pretty clever though but not new. Signals in the 10's of MHZ are typically found in modern wideband radio communication systems like Wifi, LTE, 5G, 6G. Later on the video you'll see Mr Chen shows some of these typical radio applications. But all of these work with demodulated signals in the 10's of MHz range.

[Keysight DanielBogdanoff]

Hey Dave, if we're correct about the new Keysight HD3 using the NUS ADC, do we win the HD3 as a "prize" 

I asked them about that actually, and they said they'll ask R&D. So it sounds like like there is no marketing info boasting about it.

We design our own ADCs, and usually go to one of the big fab houses to make them (unless it's one of our InP chips which we do ourselves in Santa Rosa). No InP on the HD3 to my knowledge. We do have a couple more traditional fabs, but those are used for lower volume / specialty chips.

[mawyatt]

Hey Dave, if we're correct about the new Keysight HD3 using the NUS ADC, do we win the HD3 as a "prize" 

I asked them about that actually, and they said they'll ask R&D. So it sounds like like there is no marketing info boasting about it.

We design our own ADCs, and usually go to one of the big fab houses to make them (unless it's one of our InP chips which we do ourselves in Santa Rosa). No InP on the HD3 to my knowledge. We do have a couple more traditional fabs, but those are used for lower volume / specialty chips.

We know, here's an early experimental KS Custom Really Really Fast "Stingray" 10 bit CMOS ADC

We first witnessed the amazing SiGe BiCMOS Griffin DAC way back (have images also if interested), that's how we got know some KS folks. Later designed a couple early experimental InP chips fabbed at Santa Rosa, images if interested.

KS Santa Rosa had a group of some of the best, if not the best, World Class Designers we had ever encountered, hope they still do!!

Best,

[jjoonathan]

I dunno if another one of these "inside scoop" presentations is in order, but this one for the Stingray ADC was super cool: https://youtu.be/pdbzIwelCL4?t=99

[tggzzz]

Hey Dave, if we're correct about the new Keysight HD3 using the NUS ADC, do we win the HD3 as a "prize" 

I asked them about that actually, and they said they'll ask R&D. So it sounds like like there is no marketing info boasting about it.

We design our own ADCs, and usually go to one of the big fab houses to make them (unless it's one of our InP chips which we do ourselves in Santa Rosa). No InP on the HD3 to my knowledge. We do have a couple more traditional fabs, but those are used for lower volume / specialty chips.

Back in the 80s, I knew one of the people who created those - Tom Horňak. Nice guy. (Back then 1GHz bandwidth, 6bit was state of the art)

Good to see "the real HP" is keeping up the tradition