Hi welcome to the Eev blog an Electronics Engineering Video blog of interest to anyone involved in electronics design. I'm your host Dave Jones Hi No, it's not Dja. Vu This isn't the Agilant 2000 series. I've already reviewed that one.

This is the new Agilant 3000 Series and I Thought we'd tear down this one as well to see how different it is to the 2000. Does it have just the same board? Is it just a firmware difference? Well, there's only one way to find out. Crack it open. And here it is: the new Agelet 3000 Series 500 MHz Fully optioned up W Let's crack it open.

it's about 12 Grand Worth M that new product smell and you crack it open. and here it is. It is clearly different to the 2000 series scope, which is right here. Check it out and you can clearly notice the large differences between them.

The two 2,000 series on the bottom, 3,000 series on the top. As you can see, the 2000 series is a physically smaller board and it's got this cutout which goes around like this and you'll notice that the shazzy is made exactly identical between these cuz there's the two screws up here that for the 3000 Series that is still there on the 2000. Exactly the same shazzy which they've reused, but they've done a cut out on the board. Why they've actually cut that out? I Don't really know why you wouldn't just take it across there and go down I'm not really sure because well, you just have to route that out and you don't really gain anything at the PCB panelization stage.

So I'm rather rather confused with the 2000 series I didn't really. uh, notice that before, but it really stands out when you uh, look at the 3000 Series board and the Uh input cans down here. The ad: the Uh analog input sections are physically bigger as you'd expect because they do have the uh, the Uh active probe uh input pins on them, so they physically have to have more connections on those. um.

and of course, the big change is that we've now got four A6 here, two of the new Uh waveform uh, the new Meaz Zoom 4 Asic which uh is where all the magic happens in two of these unknown Uh devices over here. Whereas we've only got one set on the 2000 series board here, so it's really is a uh, a substantial difference, you're basically getting uh, double the amount of grunt really. which is not surprising considering that the 3000 Series model does one million 1 million waveforms per second. Unbelievable.

And the 2000 series does 50,000 waveforms per second. Which given that it's probably the identical Asic is probably crippled a bit. I Think they might be able to up the 50,000 waveforms per second, but that's just a that's just a guess. And the processor is exactly the same.

On the new 3000, it looks like an identical part with an identical spec. It's the arm uh, the Spear 600 uh processor, but the Fpga the Zyink Fpga has changed. It is now a Spartan 3, Uh 1200 as opposed to the Spartan 3500 used the smaller Spartan 3500 used in the 200000 series model. And of course, as you'd expect, they've got the extra Uh channels here populated on the logic analyzer, whereas on the 2000 series model down here, they weren't populated.

Um, so yeah, maybe with the 2000 series model, they planned to have 16 channels so they designed it in and then they decided maybe it'll late stage. Well, we don't really. We'd rather have the 16 channels on the 3,000 series scope. So they just left it in there and depopulated the parts.

And as you can expect, there's minor differences between Uh boards in terms of layout because the the physical layout has completely changed. They obviously started from scratch. Really? I mean there is uh, you know, a real total difference in terms of uh, ground planes, component locations. Because as you can see, the Uh processor and um, Fpj are up the top here on the 3000, down the bottom, they're um, sort of in the center of the board on the 2000 series model.

Uh, of course you things have to be the same like the power connectors and the physical interface uh products. But as you can see like they physically move the battery on the 3000, it's in a different location. They've put these caps up here. it's just you know.

There are a fair few differences between these two boards. So they've just started from scratch. But apart from the two extra As6 here, really, there's there. Doesn't appear to be any, uh, real, You know, system level changes to this thing.

a slightly bigger Fpga, but that's it. So really, they've um, just gone for the extra As6 to get the extra throughput required. Uh, for the much faster waveform updating on the 3,000 series model compared to the 2000, but they would save significant cost by not including um, you know, these extra I'm not sure how much the As6 are cost, but they they're going to be reasonably significant. but uh, still, there's not a huge amount of difference and the analog front end will most likely be uh, well, it should be better because you wouldn't, um, pay for a 500 MHz front end and then put it in a 2,000 series.

Um, you know, I'd be very surprised if that's the same 500 MHz front end because, well, you know, maybe if they got it down cheap enough, if they've designed it cheap enough, they might be able to Share technology between the 3,000 and the 2,000 on the front ends. But you know, really, bandwidth is generally quite, uh, expensive to get. So analog bandwidth, that is, So really? I Think they probably uh, got a high performance, higher performance front end there as well. But apart from that, not much at all.

The only identical part of The layout I can see is the logic analyzer section down around here. that's the 3000 Series model and if I pan down to the 2000 series model, you'll see that it's an identical layout uh physically as far as the chip location goes. But the silk screen designators here, the actual component designator numbers have actually changed, so they've refactored that entire design when they relay out the board. That's a common practice to actually refactor it and give them all new component designators.

but really, that's the only section the PCB designers seem to have copied. And here's the rear side of the board. and as you can see, not much different to the 2000 series board, but uh, it is quite uh telling. As with the 2000, you can actually uh, see the traces and kind of work out the system architecture that they've got going here.

Now for starters, the four Asic they've got here. they've interestingly they've labeled them master and Slave so this is the master. Channel There's two Master Uh devices here and there's two slave devices over here. Very, very interesting.

Now as you can see this: uh, this is the Meaz Zoom Asic I Believe up here. And if you follow all of the Uh, all of the Uh controlled impedance uh differential pair Serpentine traces down to the bottom. Here, you'll find that it goes into all four channels so that main Megga Zoom uh for Asic or the master actually connects into the four analog channels. Now, the Slave Meaz Zoom Asic on the other hand, really doesn't uh, have any of that.

Uh, well, not that we can see actually through here cuz this is probably like an eight or maybe a 10. Uh, But yeah, it's probably like a 10 layer board. It' be at eight at an absolute minimum, so the tracers could be inside going down to the four channels, but clearly you can't see anything. There's a little bit going over to the Um Main uh Master Asic over here, but apart from that, there's not much, uh, visible.

Um, there's not much in the way of visible traces at all. and you'll notice the difference between the decoupling between the Meaz Zoom AC cup here and this, uh, well, what? I call like a secondary ASC down the bottom here and you can see this: uh, secondary ASC has a real bunch of, you know, really heavyduty, um, really serious, uh, ceramic bypass capacitors on it as opposed to the Meaz Zoom Asic which has a bunch of 100 ends. um, just distributed around like that. There really is a remarkable difference between those two and all of the logic analyzer circuitry down here goes into the Master Megaz Zoom Asic over here, so it doesn't look like any of it goes into the secondary uh Asic over here, which probably just maybe processes the extra two channels I Don't know, or maybe it helps.

Uh, helps divide up the um, you know, half a million waveforms per second each or something like that. I've really got no idea about that. Master Slave Arrangement I'd have to sit down and have a good think about it. And as before, the Master Meaz Zoom Asic goes down here, over to the Uh display connectors directly to the display.

The Meaz Zoom Asic drives the display directly. The processor up here really doesn't have much to do at all. Uh, only if it wants to do some math functions or some other auxiliary functions that are actually done in here and they're passed to the Megaz Zoom Asic which then goes in and drives the display. That's why you can get that massive display update, refresh rate.

and as with the 2000 Series board, I I'm not sure if you get this on camera, but you can see some wash residue from when they've washed the board. there's some. There's some residue left over there and that's not that great, but it's really not that much of a big deal. but um I Expected them to take a little bit more care than that.

I Expected the boards to be a bit cleaner, but really not a big deal. It's very common. There's one interesting design choice I Noticed on this board, when you look all over it, there's something that you can't see in terms of the capacitors. Now, take a good look around there, memorize it, and take and see if there's any difference on the back in terms terms of the capacitors used.

Now, if you're if you got a Keen Eye you'll notice there's one tanum there. one tanum on the entire design. Now, um, usually a a lot of companies will, uh, have like a blanket rule to avoid uh, Tanms because um, they're expensive, the material in them is quite rare. Ceramic technology is, really, um, putting big pressure on Tanms these days and uh, some of the older um style ones or the cheaper ones actually.

um, have a lot of problems uh in terms of well blowing up. Um, but I won't go into the details of it. But overall, I Guess the team who've designed this or maybe it was designed in sections I Don't know, but somebody decided they wanted one little tanum there. so that's an extra bill of materials.

um, item? Why they couldn't have used one of these large Ceramics or something like that? It's probably due to the regulator chosen, they might have to use a particular that tanem to get the ESR to make the thing stable. I Don't know, but um I Thought that was just rather interesting that they've avoided Tanms on the entire design except for one Unbelievable actually looking at the value, it is uh, 470 microfarad so that is pretty big, so really, you can't get that with the ceramic uh options they're going to have on the board there, so you know they've got one other Electro on the top side. But yeah, maybe they were forced into using the Tanum? Who knows now I'm on the top of the board here and this is the master Asic and this is the slave and you can see the tracers actually connecting the two. There's bound to be a lot more internally, that's for sure.

but really, that's all we can see. Uh, on the top side, There's not really any other connectivity here on the board at all. Okay, we've taken apart the Uh front panel display and there's the front panel PCB Very similar to the 2000 except that the board actually extends down to here here and the um, the points there. There's actually uh, pads along each input connector like that on the front for the if you have a look Bingo There it is.

They actually go in there for the um active probes, those smart probe uh interfaces so they're actually just like a digital um IO interface. There's some power on there as well I believe for pairing external uh, fet probes, but you just get that. that extra interface on the 3000 Uh series scope and really not much difference at all. But there's one interesting thing you'll knowe is the shielding on this on on.

The design of these new Scopes is excellent. and here's what they do. They actually, um, put these pads on here which are the ground. Basically, they just expose them and then they actually mate up.

When you push them together, they mate up with these uh, spring contacts down here so that you get um, excellent, uh ground shield in between the board and the actual Um shazzy. So these are these are all over the scope. They do this in a lot of places and it shows great attention to detail that they really cared about the Uh EMC When they when they design the scope, they really know what they're doing, but you'd expect that it's agilant and I really love how this whole thing fits together. Look, they've bent the Shazzy like that and that actually protrudes through through the front panel there and actually makes the Uh ground uh test hook for the Uh demo signals.

and the probe calibration signal is directly connected to the shazzy. Very strong, very nice design. and then the if you'll take a look at the board, the actual um, the actual probes are built on to the board. The test hooks are on there as well.

So when you assemble the whole thing, uh, it actually protrudes through the front panel. So these, uh, these test posts when you assemble the whole thing, they poke through there. Along with that, it's just. it's beautifully designed.

It's a beautiful example of Um system design at its best. There is a whole bunch of Uh concessions and tuing and throwing between R&D groups. the guys who design the PCB, the guys who designed the schematic, and the guys who design the housing by guys I mean guys and girls. Let's not be sexist here, and it's just there's a lot of effort which goes in into producing uh, a high-end product like this at a complete overall system level.

I Love it! I Just noticed something that is a brilliant attention to detail. I Almost missed it these two connectors. Here You can see how these two boards don't line up right. They obviously they couldn't push this connector down further because due to you know, uh, constraints that actually surround the connector and things like that.

So they they pushed it up to there for some reason and you'll notice that they're offset um in height one. This one here is higher than this one. So instead of just bending the cable, what if the PCB layout guys done, they've tilted that connector from the vertical off to the side a bit. You know, 5 or something like that.

Same here, they've tilted that slightly offset so that they line up. Great attention to detail. Thumbs up. Now, of course they may or may not have thought of that right up front.

That's why the PCB is at rev 5. Perhaps because well, they assembled it and when oops, the you know, the first prototype. oops, these don't line up. We should just, uh, rotate the rotate the connector a little bit.

And that's why this front panel display board as simple as it is, Is it rev to? because maybe something as simple as that. Um, they may have been caught out in the first prototype. Who knows? They may may not have been that smart and thought that far ahead, but kind of. You know, stuff like that can really catch you out unless you do complete 3D system modelings and mockups of how PCB designs are going to work which is one of the major advantages of today's PCB package is with all their 3D capability, you can import the model for the case for the entire external case and put this PCB in here and see how it fits.

put the other one. You can even put the connectors in there and things like that. Those sort of powerful tools allow you to get these sort of things right before you go and spin your first prototype. For now, thanks to Steve Lebson, we have an excellent Uh block diagram of what goes inside in what happens inside the new Agilant Meaz Zoom 4 system on chip Asic It's phenomenal.

Check out the capability built into this thing. Let's check it out in some more detail now of of course you got your external ad converter here that's not built into the Meaz Zoom Asic that's external and that has um four Meg samples per second pumping data from the four channels into the Acquisition memory Manager. Now these 16 Uh digital channels from the logic analyzer also fed into the Acquisition Memory Manager. Now this block here.

Clearly Um handles all of the Uh segmented memory capability cuz there's that segmented uh feature that would uh be done in here as well, surely with the help of some uh support circuitry as well. and it handles like um, uh, breaking up the data like if you're measuring two channels, it will, um, have the sample rate and put it into memory and stuff like that. So there's a there's a lot of magic that happens in that Acquisition Memory manager a lot of really high bandwidth, um stuff and that's coupled directly to the built-in Dam Now the ASC has four Meg samples of built-in uh Dam on the 3000 Series model I'm not sure if it's a slightly lower spec Asic for the 2000 Series model which only has 100K but uh, the 3000 Series certainly has 4 megabit of Dam or four Meg samples of dam uh built in. and that's phenomenal.

And having it directly coupled on the die like that, That's how they can get the Um the fast updating rate because it's going to be faster when it's on di it's more tightly coupled uh, straight into the Acquisition Memory Manager and that's how they do it. now. this goes into a display plotter over here, which we'll go into later, but there's a measurement buffer up here, which uh, presumably that would take um, just uh captures of the data and then uh would have its own analyzing circuitry to actually do your onscreen measurements like your you know, your RMS voltage and your average and all that sort of stuff would probably done in that measurement buffer. There's no other detail there, so let's assume that's what it's actually going to do.

And down here, they have Hardware serial decoders built in. This is, uh, traditionally a feature which uh has been performed inside the CPU up here, which we'll also go into, but they've decided to build it onto uh, the die itself and as you can see, they're they're simultaneous so you can decode uh, I squ C you know, SPI USB um, all that sort of stuff. All those serial Protocols are done in real time on the actual Hardware itself, which is phenomenal. That's another traditional one which is done on the CPU, but because you got a buildin, you've got speed advantages and you can actually decode stuff in real time.

Of course, they got all the triggering uh capabilities built in there as well as you'd expect. And they've got the waveform synthesis uh engine too, which is uh, presumably like an arbitrary Uh waveform capability. now. I'm quite disappointed that it doesn't actually have to the user an arbitrary capability, because it would have been phenomenal to be able to capture data, put it in memory, and then uh, store it and or modify it and then output it as an arbitrary waveform directly from the waveform synthesis engine.

So it's a bit of a shame that it doesn't have it. but I Suspect there may be arbitrary type capability in there, and with the test waveforms, they're probably I Don't think they're stored on chip here? they're They're probably stored as part of the CPU up here and the CPU down dumps the data um for the test signals into the wave for synthesis Um engine. Now as you can see, they've got masking capability as well built in another traditional function that the Uh CPU actually uh takes care of. So in your traditional Uh scope, you would have your Ad converter here and you might have a big Fpga or a custom Asic like this doing various stuff, but then it would usually funnel the data or there'd be like a Je Port memory.

It funnel it into external memory and then the CPU would access the memory and then update the display and then your CPU becomes the bottleneck and that's the real problem. So that's what they get around with with this new Uh custom Asic Because everything's done on here through the display plotter directly coupled onto the LCD display, the CPU is really relegated to just like a secondary Uh function Steve's put in here like math, measurement, and search capability and things like that. But of course it runs Windows CE and it handles Ethernet capability USB and file stuff and all that sort of thing. But really, it doesn't talk to the LCD direct if it wants to do something.

it goes through here. and there's a gooey controller as well, which handles you know, all of your uh, graphical interface type stuff. Maybe your cursors and all that sort of thing are handled directly via the Asic. so your CPU has just become a supplementary item in these modern Scopes It's phenomenal and then it'll output some data directly into other capabilities.

this Um Acceleration for measurement, search, and things like that. That may be the separate Um Fpga on the board. I'm not quite sure of how because you saw in the tear down that the Uh, that there's an Fpga and and next to the CPU so possibly that's what that uh Fpga does, but that's only a guess. But as you can see, it's phenomenal and it's entirely different to your traditional Uh scope design where the centerpiece is a CPU, But that becomes a real bottleneck and you can avoid that with a custom Asic like Agilant have done here in the new Meaz Zoom 4.

It's great I Love it and that's how they can. It probably took them years to develop this Asic but once you do, once all the hard work's done, then you can bring this capability down into your lowend. Scopes You can advertize the cost down into the low end and it just, uh, you know, allows all this phenomenal update rate capability in low-end Scopes Whereas traditionally you had to pay1 $20,000 to get a million waveform updates per second, which is what this Asic is capable of. One million waveforms per second directly onto the LCD display, there's no way you would ever do that with your traditional CPU approach.

Now that's on the 3000 Series model I'm not sure if the 2000 series model uses exactly the same Asic or whether or not it's a scal down uh version which has that less uh sample Dam and less waveform update capabilities per second. I'm not sure but uh, it certainly allows you to bring that capability down to the affordable level. It's great. Agilant! Huge thumbs up to Agilant for Designing this new Uh Meaz Zoom Asic I Love it.