Hi. Unfortunately, the Agilant 90,000 series scope here $140,000 worth has got to go back. It's only a loner, not that I' have much use for it here anyway to probably just sit under the bench here Gathering dust because it's absolutely useless as an everyday scope. It's just too high performance.

It's too noisy, it's too convoluted to use it. It's hopeless. But I Thought we'd do some quick tests before it goes back. Uh, probing a signal which really only a scope of this caliber is capable of doing I Thought we'd measure USB 3.0 super speed and see what we can do.

Let's go. And by the way, I'm not going to be doing serious, you know, in-depth measurements of USB 3.0 Here it's just playing around. see what we can come up with. Maybe get an eye diagram something like that, maybe even do some serial decoding of the data if that's possible.

I don't even know if this thing has the software capability. the software options to do that I know it's physically, uh, capable of doing it so and I haven't read the USB 3.0 standards. None of that, so don't and moan that I'm doing something wrong, not doing it right I Miss this miss that. We're just going to have a play around not reading the manuals, doing nothing.

just a play around here. see if we can get something. um, some sort of uh measurements and eye diagrams and stuff of a 5 gig bit uh per second USB 3.0 signal. It's going to be fun all right now.

What we're going to be measuring is one of these icy box uh brand USB 3.0 RAID Hard Dual Hard dual RAID hard drive uh systems here which I use for my backups and I've uh, taken it apart here so that we can access and probe the Uh signals on the back there and uh, we'll have a look at the Uh probe as well I won't have the hard drives plugged in? um cuzz I Think we should actually get uh, highspeed data in and out of this thing. Um, even though we haven't got the hard drives on there, you know it should just do some coms and uh, stuff like that. So we should at least be able to get some data out of this thing. and that's what I'm going to probe I'm going to probe the uh highspeed side of the USB connector because if you haven't seen the USB 3.0 connector, there, it is.

It's um, it is different to your standard connector. It's got two extra pairs on top here. it's Backward Compatible um with the USB 2.0 but it's got two extra highspeed pairs on the outside of the connector here and that's what's used. one transmit, one receive and that's what's used for the 5 gbits per second data transfer in addition to the USB 2.0 the Uh standard single pair up here with the 5V power as well.

So there's there a transmit and receive pair on there we can probe. so let's give it a go. and the cable of course is USB 3.0 and it's got SS on there which stands for super Speed which means it is, you know, 3.0 it's rated for that 5 Gbits per second and uh, my HP uh DV7 notebook Here it has also got um, a couple of um Super Speed USB 3.0 ports on it. Now if you have a look at the board inside this icy box here, you can see this is the US speed connector here and you can see that that uh differential pair on the top there that comes from the that's a standard USB pair there, but you can see two additional pairs there AC coupled there.

They've got a series cap in there and two other differential pairs. they're the highspeed ones which are the additional pairs on the USB 3.0 connector there. and on the back of the connector here you can see the standard four pin arrangement in there for the Uh standard USB 2.0 and then it's got two other uh, well, it's actually got five pins, one's actually ground and so it's got the two other pairs here. This is a through hole connector and I've sold it on to 0.1 in headers on there so that we can use uh, the supplied adapter cable for our probe which we'll take a look at um to Uh probe the signal without having to hold it in place.

Now I Know this is not the best thing to do in terms of Uh signal. Integrity So if we're really probing this system properly characterizing it at these sort of Uh frequencies, we really have to probe it correctly. and I'll show you a proper probe in a second. But this will be good enough for today's experiment to allow us to uh measure the signal.

So I don't have to dick around, um, trying to probe the thing properly. It'll do a reasonable job, but it won't be perfect. And here's the probe we're going to be using today. You've seen this before.

I've showed it on one of the previous videos, an Agilant 1169a 12 GHz bandwidth differential probe and it costs about $122,000 This Uh probe so it's about $11,000 per gigahertz. Go figure. Anyway, Um, it comes with several of these little adapter cables. These are little coaxes little micro coaxes.

so it's got two of those in the end. one for the positive, one for the negative because it's a differential uh Pro and you can see that marked on there. Um, and uh. there.

There are several little adapters. These have got little uh coaxes. They'd be really top quality coaxes. This one actually has its own part number.

It is E 2678 A There you go I Have no idea you can probably pay a couple hundred bucks for this little adapter I'm sure. Um, Anyway, it came with the Uh kit here and it has a little um, uh, a little uh two pin, um, head head on there which comes with a converter to convert into a 0.1 in Uh header so it allows us to probe that directly. And here's the other probe it comes with. and if you really want to probe this thing properly, you would use this to probe directly onto Um test pads or directly on the Uh pads of the IC or the Uh Terminator that you're um using on your differential Um pair transmission line and you know if you really want proper signal Integrity this is what you got to use and we won't I Don't think we'll bother with this today.

Once again, it's got the same little micro uh coax interface, but it's just got uh test pins that allow you to directly probe the circuit under test. And for those who really want to know, we're using an As Media ASM 1051 chipset. All right. So I've got my probe hooked up here uh to one of the lines I'm not sure whether whether or not it's transmit or receive, but this one seems a little bit cleaner than the other one.

So I'm going to assume it's the transmitter because it hasn't. Um, you know the signal hasn't degraded by the time it's got all the way down that 1 M long cable or or whatever it is. so this is just you know, your basic regular oscilloscope. Um, uh, you know it's triggering right in the uh oop Center of that Bloody touch screen's really annoying I shouldn't poke at this thing Anyway, we're talking 200 M volts uh per division there.

So we're talking 200 400, 600. you know, 7 or 800 molts Peak to Peak there for this differential signal and uh, as you can see, it's just all that random data in there. and we're looking at 500 ponds per division there. so each one of those you know you can see one cycle there is.

You know that's 200 PCS per division. So it's incredibly quick this thing. but this is a 13 GHz 40 gig sample per second scope. and you can see that it is saying that we're operating at 40 gig samples a second using 2K points of memory.

and uh, obviously the lower we go, it's only 1K points memory. so that's the minimum it's going to use is 1K of sample memory automatically adjusts that sample memory at 40 gig and if we go down, it's still 40 gig at 2K points, 40 gig at 8K points 20K points. So it can still do that sort of. You know it still can do 40 gig sample second at two Meg points at 8 Meg points, right? We've got a massive and then it once it goes to 10 Meg points, it drops down to 20 gig samples per second.

Oh gosh darn it can't keep up. Oh what a HP of crap. Yeah right, it's $140,000 Bloody Scope. It's brilliant.

Okay, so you're still getting. you know you can single shot capture. That single shot captured it and you can zoom in on that 8. Meg of you know it's pretty crusty now cuz it's right down in there.

but you know it's uh still, it's quite. It's quite remarkable that sample rate and that amount of memory. But anyway, we're right down there and we'll be able to have a look at this data now. Of course you know you can't tell much from this at all.

So what we're going to do is try and get an eye diagram. So uh, haven't read the manual on this thing I haven't used one of these uh beasties before so I'm going to go up in well, hang on Now let's see if we can get some uh persistence display on there and uh, let's have a look at that. So let's go into uh measure and now let's go into setup: bloody touch screens and uh display. Let's have a play around with that aha color grade.

There we go. let's switch that on and see where. Hey there we go. We've got some persistence now that you probably used to seeing on these high-end Scopes It needs time to to actually get the data to acquire the data and we getting some persistence there.

So the red stuff is where you getting more of the signal. Okay, it's right in. uh you know it's capturing that more times and down to like the green stuff there which doesn't H which only happens very occasionally and uh, you can and it's got stats down here. There you go.

It's got stats. white is biggest and uh, like to how many uh times it's appearing on each individual capture. so green is the lowest right? up to White which is the highest intensity. You probably can't see the white in there.

oh it's You know it's smack in the middle there, but there you go. That's the persistence display. I Mean it's operating very slow because it's got to do a ton of stuff. It's got 803 points there, but that's the persistence display.

but still, that's not showing us a huge amount. What we really need to look at the Uh signal Integrity of a differential signal like this is to look at what's called an I diagram. So let's fumble around here and see if we can get an eye diagram which will allow us to analyze serial data. Uh, Mass Test measure I Pattern There we go.

There it is. we should be out, you know. And it's got a ton of stuff that we can look at and we could spend days and days and days here analyzing this signal. In fact, it would probably take you a couple of days to set up the measurement if you're doing really critical.

If you're developing a USB 3.0 product, you really had to get critical measurements to make sure it past the standard and all that sort of stuff. Then you know you could spend days setting up and probing this just to get your one, uh, you know, set of measurements or something like that. So really, you know I don't have the time to uh, do that. So we're just playing around today and the USB standard would have like, um, how? uh oh.

Actually, that's just measuring. Okay, ah yeah, we're just in the measuring menu here, so we're not. uh, and we can trigger on all sorts of stuff. Let's go into analyze.

Let's go into serial data here. And and hey, here we go. Let's try serial data wizard. That sounds good.

We can do it all manually. Um, serial data analysis. Here we go, we can turn on the real time eye diagram. Actually, let's just switch that on.

It should do that in real time and at least points must be acquired for the Infinium to recover the clock. Because it's got to recover the clock. Because we've only got one signal going in here, we only have the data going in, so it's got to recover the clock signal from that before it can get that. um, before it can get that eye diagram.

So let's go into this serial data wizard see what it can do. Looks like it's going to take us to clock recovery threshold. Um, time interval measurements, real time eye display, clock, and acquisition. So let's go in here and channel one.

Yep. Second: uh, select the clock recovery method it's got PCI Express Fiber Channel Flexray. Um, so let's leave it on it. Default second order PL and see what we get.

Um, autoscale the vertical. We don't need to do that cuz it's already pretty good there. I Think so. let's go next.

Nominal data rate is 5 gbits per second. Um I believe that's correct for USB 3.0 So let's uh, enter the damping fact damping factor for the PLL 707 that'll do I'm not around going to use defaults or maybe somebody has set this up before me and they've had a play around turn on a time interval. error measurement relative to the recovered clock. Sounds interesting.

We'll turn that on and uh units in seconds. Next, and we go to the real time eye. Here we go: turn on real time eye diagram Bingo That's what we want and that's what it should look like. It'll be a bit.

it'll be a lot fuzzier than that. I Suspect because our probing is not. uh, perfect. use color graded display um yeah, why not? and uh, which bits to include? Oh God Anyway, sample rate: 40 gig sample per second.

fast update sounds good. uh main time base scale: 40? whatever. Yeah. I Just want to hang on display clock the main? Yeah, uh yep.

let's do it. finish. Woohoo! Hey, look at that. We have our eye diagram done.

Beautiful! Now this thing takes a while to process, so don't expect instant results here. And what it's doing is recovering the clock signal from this data and then it's overlaying that data on there to give you what's called an eye diagram. And as I suspected, this one's pretty fuzzy. It wasn't um as good as the example that we uh saw back in the setup menu there and it tells you how many waveforms it's captured.

There we go. It's a th000 now and um, we're at 100 Peak seconds per division, folks. Check it out. So that one that half cycle there is 200 P seconds.

And anyway, the eye diagram allows you to show um, basically you want the widest eye possible, that black bit in the middle. You want that to be as big and as wide as possible because what this is showing is a whole bunch of stuff. It's showing the uh Jitter of the signal uh, mainly which will close the eye in this x Direction like this and then you can close the eye in the Y direction and uh, the USB 3.0 spec or any of these high-speed uh signal specs will specify how big that eye has to be in uh, both the horizontal and uh, vertical directions like that. So in terms of Jitter So if um, you will get in, you know if you had a very poor quality clock on this thing on your system for for example, and you got lots of Jitter on your waveform.

When the waveforms overlay themselves on there like that, you see how there's some data points falling in the middle, you'll find that that eye will close. You know it won't be big and wide like this black. it'll close like that because your Jitter will be too your clock. Jitter will be too much and it'll what's called close the eye in that direction and that's bad.

And there's other stuff like um, uh, symbol, interference and stuff like that. You can go Google all that sort of stuff and you can spend uh, days just looking at. um, you know how these uh, what, these eye diagrams can tell you and stuff like that, But there you go, that is. That is fascinating.

That is the eye diagram. Can we turn off the color gradient on that display color grade. There we go. There we go.

So we've turned off our color grade and we can see just the raw data there, shaking around and jumping and uh, but occasionally you're going to get some data in the middle there and that's why that uh, color um, gradient is a good thing that that persistence measurement. But there you go, that is. and of course we can change the vertical on that as well. But that's the eye diagram for what I believe is the transmit signal.

so let me just probe the other one. I'll disconnect it and oops and we'll reconnect it it here. There we go, that's the other one. Hang on.

does that look better or worse? That's a bit worse I think it's a bit worse. So and yeah, there we go. That's got issues with uh overshoot there. you can see the so that is clearly the the receiver on this icy box being because we're getting Um level issues there due to our nonperfect transmission line.

All right, What I've done now is I've um turned on the RMS Jitter measurement. So it's attempting to measure our Jitter in there and you can see those Auto cursors set up at that point. and there's our RMS I Jitter at mean in is 1.27 Picos seconds and if we uh, go back in there and do that again and say choose Peak to Peak instead there we go. It's much larger now because it's getting the peak to Peak value of that Jitter You can see basically from the Um opening of one eye to the other eye there, the distance between those two black points in the eye is the peak to Peak Jitter and that's uh, where is it? Um, we're talking 197 P or 86 PCS or thereabouts.

And there's a whole bunch of other stuff we can measure. we can measure you know duty cycle Distortion Q Factor Crossing and the height and the width of the eye and stuff like that So you know if you want to measure the eye width for example, use measured data, extrapolate, use using standard deviation and really, you can go to town on measuring. You know, uh, you know, just measuring this stuff correctly and let alone probing the thing I Mean we haven't even probed this thing correctly, let alone measuring it. Um, to the exact specifications of the USB 3.0 standard I'm sure it's very complex I Don't think I've ever read the USB 3 standard actually, but I'm sure it contains all of this stuff which would be, uh, critical to get correct signal integrity.

And if you're designing a USB 3.0 product, you really need to get this sort of stuff. Um, right. and it would you know, be over a standard length and type of cable and you know into a specific load and all sorts of stuff. So um, you know, really? you could probably spend weeks actually characterizing your system for to make sure it meets the correct performance targets.

I'm just by Maring around with the clock recovery here. and really, if you don't get it right, you're screwed. I Mean if we set that to PCI Express for example, we're just. you know, we're just not going to get anything there.

and if we set to first order PLL nominal data rate see at 2.5 gig bit, we' got nothing at 5 gig bit there. There we go, we are recovering a clock there. So second order PL which is what we were using and uh explicit second order: PL clock frequency 5 gig? yeah Loop and with got bugger all there. So there we go wh he doesn't like that 5 gig There we go Bang where spot on there and I've got um, we're measuring the transition uh bits there and that's the D emphasis bits and what we were looking at before which was both combined.

There we go Bang! That's the usual way you want to do it and we can do Mass tests and we can set up. You know we can set up automatic Mass testing which you've seen on like the cheaper Agilant 3000 uh X Series uh Scopes and stuff like that you can set up um IE diagram mask test but it wouldn't be capable of doing us 3.0 cuz the bandwidth um just isn't high enough so you know we can turn on mass testing. but I'm not going to muck around with that and it just tells you if it passes, passes or fails a preset up condition. um, for your mass, you know if your eye is too small, it'll automatically tell you.

It'll tell you how uh, how often it fails. You know how many times per second it fails and whether or not that meets the standard blah blah blah. and uh, if you want to get um, your USB uh three product actually uh, you can send it to a lab and they will measure. They'll know how to do all this sort of stuff, know how to analyze it, Use these instruments like these and they'll uh, provide you a report telling you if your uh product passes or fails the USB 3.0 spec and we'll try and, uh, see if we can maybe do some serial decoding perhaps I'm not sure if this one's got it, but um, setup.

Jeez. these touch screens are really touchy pun intended. Um, serial decode. Hey, there we go.

Hey, it's already on USB 3.0 And here's some of the stuff: USB 3 2 SBI Sarda um PCI Express Gen 3 all sorts of stuff JTAG Lin Infin Ban Flex Rid Ethernet um 10 gig. it does 10 gig. um uh. ethernet as well.

But where want USB 3.0 show decode? is that that's grade out. So data s Source One Channel One bang uh descramble Show decode. It's not. There's no Auto setup.

it's all. uh, it's all. it's all gone. It's not going to let us.

uh, do it. I'm not sure what's going on here now. this looks interesting. I've gone into the uh Jitter analyz Jitter men up here and where looks like we can get a whole bunch of uh, various Jitter measurement stuff.

So this looks really quite neat. Data Time Interval Data rate H Clock Recovery rate D emphasis It's all there time interval error and let's select something. Jeez, we could really go to town here. it's not given us anything.

Maybe we don't have, no, Maybe we haven't started. it. probably haven't set it up right. Anyway, it does.

It will allow you to do all sorts of bit error rate stuff. Ah man, you name it, this scope can almost certainly do it. I I'll try the Jitter Wizard here worked for us before with the eye diagram. So let's uh, select measurement.

I've selected the Uh data rate. So Data Rate: Fully automatic. Channel One: Yeah, let's go in here: Thresholds, individual sources Channel One: Yep, whatever. plots a histogram of all data rate measurements.

Yeah, this is what we need. Turn On Measurement Histogram: You bet it's not automatically showing up. Yeah. Jitter Spectrum Turn on the Jitter Spectrum You bet.

Finish: Hey, There we go. There we go. That's a histogram of our data rate. Now where's the scale for that measurement? Histogram Data Rate: Uh, there we go.

We're actually at standard deviation. Where uh, sorry I mean is 7.4 gbits per second. What's going on there? There it is. Mode: 5 GBS per second Mode Minimum Maximum Hey I don't know what's going on there, but uh, there you go.

We can actually get up. Ah, measurements. There we go. Data Rate: Yeah, where it is actually telling us that we are at 7 gbits per second.

It definitely has to be something wrong there cuz it's showing. Our minimum is 3.7 our Max is 12.8 gbits. there's something something screwy really going on there. So I don't know.

I'm not going to spend time mucking around trying to get that sucker working, but you see that it can do neat stuff like that and much much more as well. I'm sure. Well, as for this uh, serial decode thing I relented and I called up the help for USB 3.0 and it says, you know I show Decode I presume. Um, this scope does not have the USB 3.0 serial.

You know it says it requires the Uh USB 3.0 software. Um, so I can only presume because it's grayed out. Um, these things were all grayed out It didn't Um, it hasn't got the software option built in for that. Bummer.

I'm sure that costs, uh, thousands of dollars when the scope costs. uh, $140,000 So no doubt you got to pay for all these individual software options. So yeah, I think it's just. uh, it's just not going to let us.

uh, do that. There's no Auto Setup? No, nothing. Um, so looks like it doesn't. It doesn't have it.

Show Decode can't do it. What? What? I've called up here is one of the Uh demos for the real time eye with mass testing and it's put I can close this stupid screen. There you go, It's it. Showed us a really clean eye.

Check out that. I Mean you know, check out the amount of black pupil in there for want of a better term. you know, um, it is. You know that is a real clean differential signal that's a beautiful eye diagram that just you know if you see that sort of stuff on your if you see that sort of waveform on your system, you know you're You're really sitting sweet.

You're laid out your board properly. You got all your controlled impedance traces correctly, All your termination right. You know all your shielding's right. you've got low clock.

Jitter Because you know, look at the you know the width in there is just. ah, it's just beautiful. So this is a demo. Um, if you probe the thing properly and design your system properly, you might you know you might never achieve that in your system.

You know it's all about meeting the specs and the tolerances and what is actually achievable in a practical system. But that uh, I don't know if I can do anything else with that demo. I Think it just uh, displays. Um, it was just a demo file really.

And it allowed me to uh uh, maybe I can do here we go. USB 3 Serial decode. Maybe this is what it would it look like? Um, if we were actually able to do that. Yeah, there we go.

This is. This is what we would have got if we had, um, were able to get that serial decode option working. There's the USB uh packets, the time, and all sorts of stuff and you can analyze that until the cows come home. Good stuff.

More neat stuff this can do is uh, Fpga Dynamic probing and uh, this. this is the demo of it and if you insert a Uh core a specific core into the Fpga and you probe it, um, then you can get all sorts of stuff. It helps us set up your scope and get real time analysis of what's Happening inside your Fpga Neat. Sure it cost a Fortune of course.

So there you go I Hope you found this rather interesting. If you want to check out more, um, go search for you know Google Uh IE diagram uh measurements and stuff like that I'm sure there's 10 million app Notes and uh things out there which will, um, explain just what a valuable tool this is for um doing signal Integrity measurements on differential um, highspeed serial lines like this, and um, pretty much on all you know these high-end Scopes This is one of the main things that they are capable of doing is uh, measuring and analyzing and uh, you know, extracting the clock from the uh Uh signal and display in Um and Ie diagram which is Um essentially the data waveform uh, overlaid again and again and again on the recovered clock which Um controls the horizontal Um axes is the recovered clock from the signal itself, and from that you can tell an awful lot about the signal. Um, but it was just good fun. Um, having to play around with this and see what the scope is capable of I Really rather like it.

It's been fun and uh, if you get a chance to play with one of these, uh, high-end Scopes by all means, Um, play around with the Ey diagram stuff because it can be a great deal of fun and you can learn a lot about signal Integrity You know if I could I could, uh, you know, probe this in different ways and uh, stuff like that and see how the signal Integrity changes. but uh, haven't got the time. So if you want to discuss this, jump on over to the Eev blog. Forum The link is below the video there or above if you're on the blog website and remember if you like it, please give it a big thumbs up.

Catch you next time.