EEVblog
Вставка
- Опубліковано 12 вер 2024
- Inside the National Instruments NI VB-8034 Virtual Bench. A 350MHz 4 channel mixed signal oscilloscope, arbitrary waveform generator, power supply, and I/O module that can connect to the PC, tablet or phone via USB, WiFi, or Ethernet.
www.ni.com/virt...
Datasheets:
Xilinx Kintex 7 www.xilinx.com/...
ADC08D1520 ADC www.ti.com/prod...
CRF Reed Relay: standexelectro...
High Speed OpAmp: www.ti.com/prod...
Forum: www.eevblog.com...
EEVblog Main Web Site: www.eevblog.com
The 2nd EEVblog Channel: / eevblog2
Support the EEVblog through Patreon!
/ eevblog
EEVblog Amazon Store (Dave gets a cut):
astore.amazon.c...
T-Shirts: teespring.com/s...
💗 Likecoin - Coins for Likes: likecoin.pro/@...
I actually had the pleasure of designing a few of the power supplies inside this guy a few years back. Pretty cool to see something you've worked on in a shipping product!
+RadRider33 That's really awesome.
+RadRider33 Small world!
The video ad that played before the start of this video here was for this same NI VirtualBench product. For those that haven't seen the ad, it showed how a light bulb improved upon the kerosene lamp, the washing machine improved upon hand washing and that the NI VirtualBench will be the way forward in the lab. If the 4 digit price just needs a decimal point placed in the middle and I'm sure it would... ;-)
@23:00 I've had so many bad experiences using built in ADC's in MCU's that I'll never do it again. I stick to SPI ADC interfaces for resolution and isolation.
Pretty impressive quality, but I have to say with that price point I'd just go out and get some Rigol gear, a couple of other goodies and use the rest to book a vacation at the beach.
Good idea. Instead of hiring you, ill hire the chinese and take a vacation while you are unemployed.
+jak p (skiguy09) sounds like a plan
+The Current Source You do realize that the rigol gear would not come with "The Smell" right? I dont know how you would get anything done without that.
Mark Webcraft
Ah, major oversight on my part.Thank you for pointing that out! :)
+The Current Source After paying around $4k on Rigol 350 MHz four channel scope and buying couple of other goodies I don't think your vacation will be too festive;)
At 30:00 "496 dollars. US of course" And the image says "All prices are in AUD"
lol dave :)
+Hene193 DOH!
+EEVblog so which is the right currency?
+william montes Pesos.
Toni Lähdekorpi Thanks
Lets put a V8 on a roller-skate, Nah, lets put TWO V8s on a roller-skate.
Heatsink sense probably to avoid cooking the expensive FPGA - minimal cost to implement so makes sense
+mikeselectricstuff my guess is more than one blew up before they added that.
Yeah, those Kintex chips can dissipate a lot of heat if they're "full" and running at a reasonable clock.
+mikeselectricstuff Given all the test pads on both sides of the board, my guess is that this thing is production tested without the heatsink. Of course you don't want it to overheat while verifying it works properly ;)
+mikeselectricstuff Kind of a huge and obvious manufacturing step to miss though, but yeah, nothing wrong with over engineering.
+iamjadedhobo Could very well put itself in a lower power test mode. In fact the heatsink sense could enable the test mode.
They must have written the whole software and HDL stuff in Labview so they need this big-ass hardware. ;)
Those look Exactly like PTC fuses... Just not the usual green colour
+mikeselectricstuff can confirm.
+mikeselectricstuff Yup. LF = Littelfuse.
Next to the "Heat Sink Detect #4" arrow, there is a number "4" cast in the heatsink itself, which refers to the screw. So it might just be an assembly hint (like the power supply binding post colours on the front panel PCB). Or maybe the detection is done by testing electrical continuity between two screws.
Hello.
I have used NI USB-6251 DAQ in the past. My speculation for the extra FPGA processor is to run locally, instead of of of your computers slow USB connection. Allowing all processes to take place simultaneously without any real-time lag between control and read-back of data. I like this tear-down...
Thanks,
Tom
Exactly the sort of equipment that was bought on my high school.
Overpriced stuff, that no student was able to use. I think the main intention of the teacher was to spend the budget, even if we just measured the characteristics of some transistors and diods. Wasted money.
Giving the students a analog oscilloscope most of them would have failed when they have had to change the timescale.
But for sure there is some AUTO button that does this job for you.
"smells like a top quality instrument"
Uh, Dave, are you feeling okay? Have you had any head injuries recently?
If the price was lower or country of origin different, surely more negatives would have been found!
Nice teardown... Now turn it on!
The "Blast Shield" around the DMM is actually a shroud for the temperature controlled oven to keep the DMM at a specific temperature regardless of incoming air temperature or exhaust fan speed. The heat sink is large because the scope and ARB are also temperature controlled.
I didn't see any heaters on the video, did you?
+MFeinstein At 5:55 and 17:00 there's what looks like a row of 1/4W, 1206 resistors on the inside of the shield but the outside of the isolation. You can see it again at 19:57 topside and 20:29 bottomside.
+The Free Parts Bin that's pretty common on DMMs, Dave even explain them rapidly, it's a cheap way of making a high voltage input resistor... The total voltage will divide itself in each resistor in series, so you can use 10 resistors that have 1/10th the maximum voltage limit each, instead of one resistor that will need to take the whole high voltage itself and usually are pretty expensive.
+The Free Parts Bin Too substantial for a simple airflow shroud and looks like it could be polycarbonate .
+MFeinstein I wasn't talking about the string of resistors inside the EMI cage. I was talking about the straight resistors outside the EMI cage.
"Thermal Sponge" would be a great band name.
+simontay1984 Hard core Electro?
+enlishbob HOT core Electro
Instant 90's nostalgia with that mention Pete. There is a band called Sponge, and I think we could re-appropriate the lyrics here:
♫When I tear-down, so much compound, tiny little screws, I've had it up to here..♫
ua-cam.com/video/L65NNh6vJ_Q/v-deo.html
29:57 Dave: "US off course. Yankee money". Digikey:"All prices are in AUD".
I think I saw some simalar units to these when I was at UVA for an open house, I thought they were a bit cheap and crappy compared to "real" tools, like trying to to bundle everything into one multi tool unit. I'm glad to see those things are actually really well built!
Haha, now I am at UVA for EE, and get to use these every week. Software is a bit clunky, disappointed in it, given the quality of the unit itself.
Outrageous price for that unit. I'd rather buy another jet ski. LOL
don't think you can make a jet ski count as a bussiness expense though...
Do you think they will overlook the cocaine and hookers?
PS I bought one of 'em lunch.
Very nice device !! and Dave you did a very nice teardown ! I love it the way you do that :-)
Greetings, Mark from the Netherlands
Those thermal pads are not designed to be adhesive, it's just a side effect of the amount of compression they allow for great contact.
Notebook graphics cards use them on the VRM portions of the heatsink a lot too.
Really beautifully designed.Thanks for the teardown.
Why didn't they make it "modular"? With multimeter as it own board, that way if a student fucks up it's much easier to fix. Just replace the burnt board?
+Albin Ekblom A Modular design cuts into the profit margin.
Anatel's decal on the back. Looks like NI has its eyes on Brazil.
31:10, No. This Zynq is not 150 or 200$. It is about 50$ tops. More like 30$. I do have boards, with custom made complex chips, AND Zynq high end dual core ARM and fpga fabric, and it is all 90$ total (with "profit")!
Just a guess, but I believe the plastic tray around the multimeter fuses on the bottom is to protect students from sticking their fingers further inside the case. That's the one part of the case they're likely to crack open.
+kieferonline Maybe, but it makes no sense having the same thing on the opposite site where noone could reach the pcb without opening the case.
I think I took a BNC splitter from the lab once back at university. Don't think i'd be bold enough to nick a whole bit of kit.
Gilding Lily factor (GLF is high with this one... I believe is exponentially grows with each unit of $1000 > $2000. For an all in one multimeter-scope..L_g = x_0(1+r)^g - done the math..theoreticslly.this one would be audiophile quality as well.
i have used kintex FPGAs before ... for 10gb SFP+ transfer of PCIe acquisition cards
what the hell are they using it for in this?
+Alyx BioHaz It'd a lot'a FPGA that's for sure.
(18:48) I think the heatsink "detect" points are marked on the heatsink itself, because the PCB label matches the "4" engraved on the metal.
I can't tell if they are testing points other than #4, though. I didn't see connections to the other rings on the PCB (although they could perfectly be in the inner layers I guess).
I realize this video is kinda old now, but great teardown as usual, Dave! I can't believe I missed this one.
This entire thing, despite it being well-built, appears to be using more expensive kit for the heck of it. I am sure for what it's intended, they could get away with cheaper hardware while keeping the same build quality and results.
+bluephreakr It's intended for the educational market, where outrageous prices for hardware are the norm.
Worst part is, after a few months in a classroom, these things will be abused and beat to hell.
+bluephreakr They could, but that's traditional National Instruments for you.
thieves usually don't care how the money gets spent
I know NI has had custom chips in the past. I don't know if they still do. We have the 2-channel version at work and love it! I'd love to see how the insides compare to the 4-channel.
What an impressive and badass instrument!
"all the electrons are gonna fall out"
I thought this was gonna get old. It hasn't :D
Can't wait for the review!
Very nice tear down. Would be interesting to see some comments from NI about the huge overpower inside this instrument?
I understand the professional layout and insane build quality, this ends in an expensive price of shore. Nothing wrong for a instrument for school. But why put this lot of power and memory (money) inside and don't use it???
This channel grows great and I love every video. Thanks Dave!
Loved that tear down Dave.....
Looks like that big heat sink with the thermal sponge under it forms its shapes when its applied from new, I bet Ti will have to replace it before it goes on sale to ensure it will mould itself properly. I think to get it to sit right after removal will be difficult.
+Michael Hawthorne Agreed - I wish him luck getting it properly back together plus he'll have to replace the phase change pads on the FPGAs.
+Michael Hawthorne Yes, thermal gap pad such as this is typically designed to be a one-time use or it can deform and not provide the proper compression.
The good news is one day these puppies will be on Ebay for hopefully a great price.
***** 8-10 years is my guess. Better off buying dedicated tools.
the Kensington lock will not protect it from being nicked by the students but from being "borrowed" by your colleagues...
Love the solid screw mount surrounding the USB port. No board flexing as persons jam USB plugs in with unnecessary force. They really have designed these to be abused, haven't they!
I want this for Christmas!
Some are scared of the price but can you really get a comparable mixed signal scope, 5.5 digit DMM, digital IO, and small 2ch PSU of that quality for any less? Doubt it. This is not a Rigol. The hardware is high-end and I could only imagine the software is likely pretty good.
+JetNmyFuture
Definitely can't fault the quality.
But I have a few issues: a "normal" lab power supply and a good DMM can easily be had for less than 1k $ and, if not treated very badly, are useable for a decade and much more. *What happens when NI decides to stop supporting the workbench with its software?* A perfectly fine PSU and DMM hardware, but no way to use it...
That leaves you with 5000$ to spare, which will get you an equivalent brand MSO+fgen from Keysight/Tektronix (and I would even consider Rigol). Again, a scope will work decades, if treated well, and no missing driver/OS support will suddenly make it an expensive brick. So I would never recommend the workbench over dedicated tools.
From an *educational standpoint* (I've been teaching electronics to students at uni and for the courses it had been decided to use NI Elvis and myDAQ, which you could see as the low end version of the workbench): it is an *advantage* that all the hardware can controlled by the software, so it makes some tasks, like bode plots or tracing component characteristics, very easy. But the *disadvantages* IMHO *outweighed* it. I don't really know why, but the knowledge of how to use the tools in software did not transfer over to use the real tools. Give the students a real scope or DMM and they were stuck. There's also the quirky issues that sometimes the software has bugs and certain settings give wrong readings. And there's the annyoing issues with hardware not being recognized by the PC, sometimes licencing issues, driver issues after the IT department updated the PCs aso.
The price in Sweden is > 50.000 SEK in $ 6.250.
Should have been gold plated for that sum.
The Kintex devices have much faster deserializers than the Zynq devices. Maybe they use the Kintex for receiving the many Gbit/sec from the A/Ds and the Zynq so they get the networking and seriously fast ARM cores. The Zynq-010 devices can be had for $50 in the cheapest package.
Nice Breakdown!
Hi Dave, at 36:39 that is an LMK0482x Ultra Low-Noise JESD204B Compliant Clock Jitter Cleaner with Dual Loop PLLs from TI.
What! Where's the GPIB? What kind of NI product is this? The heatsink on the o-scope board reminded me of a telecom system I once worked on. We had single heatsink spanning multiple BGAs, but in our case, the bottom of the heatsink was milled to match the contours of the components. Since there's still some variation in the final assembled height of components after soldering, we used a two-part thermal adhesive to attach the heatsink to the board. The only way to get the heatsink off was to soften the glue using a heat-gun. Expecting something similar, I thought the heatsink was going to win this time. Cheers from the SIli-valley!
They should put a 15-19" LCD screen on the unit (whatever the size is), maybe a flip-open solution like a laptop. Would make it massively more practical for little added cost.
Always dreamed about one of these.
Power factor correction built in? Then as a bonus this will pay for itself after a year. Just leave it plugged in and you'll save money on your bill!
good call on the comparators 41:10
SNx4LVC541A Octal Buffers/Drivers With 3-State Outputs
Great tear down and a great piece of kit. I would also like to see what the current version of the software is like - if you get the chance.
@22:00 "Analog Devices AD4610 Very Schmick Low-Noise Low-Everything Super-Duper Whiz-Bang Bob's Your Uncle Operational Amplifiers" LOL Dave, imagine that being the product description on the datasheet :D
+MrSergeng National Semi's datasheet for the LH0033C and LH0066C was headed 'Fast and Damn Fast Buffer Amplifiers'.
damn the aesthetic design is really nice
Not a single bodge wire on the board? RF shielding everywhere? Top class high quality FPGA? Probably one of the most beautiful designs I've seen in a long while.
Did anyone else see that PTC fuse around 10:30? Massive! Never seen one like it.
I was waiting for this!!
looks awesome!
Our school is in connection with National Instruments and they gave our school 7 of these. We are not allowed to use it(first year) but as far as i know the second grades are not allowed too.
+Gerald Musy Probably :D
Hey Dave! Loving the teardown especially!! I'm wondering why, at 43:09 when you are looking at the double sided thermal tape there is a place for a TQFP chip that goes UNDER the heatsink! ?? Why would that be there??
Well there you go Dave, that's why you *always* turn it on first, THEN take it apart. How else are you going to know whether or not taking it apart is what screwed it up?
+Keith Ward He has confidence in his skills ;-)
+Peter “Crackpot Pete” Carlson I have confidence in his skills too ... what if it is bad out of the box? Just saying it's good to have a known reference point.
Keith Ward
Believe me, they have enough confidence in his skills to NOT want to send him a bad one.
you can double that price if you want competent software, the stuff that comes with it is undoubtedly gimped beyond use, and the SDK is designed to be as difficult as possible ... just like ever other NI product I have ever used. Which is a lot
I'm surprised no mention was made of the board-mounted threaded inserts. I've never seen those before.
Good teardown.
my guest is that the heatsink detect #4 is just to assembly people make the number 4 on heatsink be in the right side
It have a sense wire. I think the big chips will melt without heatsink and hope not power up without sensing.
I don't know how I feel on this thing. I've seen lots of blown up equipment in the labs at uni but now you also blow up 4 other things along with what you muffed up.
+Fan119 i am not sure what specific use case they are targeting this for, but i doubt they'd make it if their market research didn't indicate a desire for it... but hell, i've been to a couple high schools that could probably swing 5-10 of these for advanced classes...
+Fan119 We had NI's PXI setups in our labs and students would blow cards out of them all the time. Replacing the entire thing would be a financial nightmare.
+SuperAWaC The market for it is educational institutions with broken budgeting and buying processes. Wouldn't be surprised if they ended up sitting in a cupboard somewhere, broken.
Price point seems high till you consider how it will integrate easily with ni software such as labview or test stand.
Can imagine it'd be great for school
I love NI ELVIS II+ workstation. Its like a giant Arduino with 100x more functionality. I wish i could wonder out of the university`s lab department with one in my lap.
Minute 28 I saw it coming: NI vs Dave 1-0 Fortunately, nothing can resist a good screwdriver in a steady hand. Good on you, mate.
Did you use some thermal paste when you put it back together ? Maybe a teardown of an UPS. Thank you for the video ! tjl
Surprised to see this kind of high caliber design in an educational product. I assume this is made for beginners in electronics because advanced students would likely be supplied separate instruments for each of these purposes.
If LabView weren't so terrible, I might like NI more. Their hardware is always nice, but it's almost always been easier in my experience to just use the C bindings their drivers provide. Coding data acquisition and processing in assembly while getting a root canal would be easier and more pleasant than working with LabView on anything of even moderate complexity.
+SAI Peregrinus Someone always has to point out how another programming language sucks. LabVIEW is just another programming environment that allows quick and easy user interfaces to be created to interface to instrumentation or other hardware. Every language may do something better than another, but there is really is no one best language for everything. It is more of a left or right brain preference to either a procedural or graphical programming environment.
+Keith Ward Certainly, that's why I phrased it as a statement of opinion and not objective fact. I would dispute the left/right brain aspect, and that it's linked to being a graphical programming environment. My issues with it don't come from that: it's basically a functional programming language (which I like) but without many of the important features of modern functional programming languages (typeclasses and other type management, monads or similar, better debugging). I don't deny that it's a fast way to make a UI at all, it just makes complex logic harder to manage than necessary.
at 33:24 Dave goes on about the Zynq and the larger FPGA not apparently having much interconnect but it seems unquestionable that there'd be a ton of that on the inner board layers; your're routing two large BGA FPGAs together in a piece of test equipment, pretty much the only way to go. That thing really does have a fuck-ton of FPGA action going on.
NI: hardware 10/10, software -1000/10. Luckily there is the perfect support from within MATLAB/Simulink, good libraries to go with C/C++/.NET etc. that makes the investment somewhat more worth the horrendous price of these products (NI that is).
If I had the cash, I'd be buying one of these.
I'd be more impressed if the thing didn't cost 6k. They say an engineer will do for 5 cent what an idiot will do for a dollar.
+lodevijk Actually, I was wondering what the price might be, so I went "let's see, think 'ludicrously overpriced' - say, $5000... hmm, good start but make it more realistic in its ludicrousness - say, $5700..." Dang, I hate it when I'm right about prices...
FYI, just the MSO part of the Vbench is ~10K from competitors. This is actually the lowest cost 4 channel, 350Mhz scope on the market.
It's obvious they have chose the component parts to justify the sales price, without even fully implementing their capabilities.
if you could you please take a look at the NI DAQ? It comes in an educational format with a FPGA.
Very very nice. At 38:20 there is an A71L chip in mid field that looks like three of its legs are unsoldered! Odd.
+mbaker335 And what's with that solder splash(?) around the bottom-right 4 pins of that National Instruments chip?
+mbaker335 It has? The A71L looks fine to me.
+KhronX Looks like it is "repaired". Bunch of small components not reflowed 100%, others appear cracked (C750 on bottom)...
zlotvorx Hastily, perhaps - that thing could've used (way) more flux there :)
KhronX I have a feeling that it's made just to fill a market gap. For sure there is a way to improve it, and in the same time it leaves place to upgrade if needed (or payed for).
Is this thermal conductive foam material reusable? Because, it seems it stays in its compressed form, once used, and there may be not enough mechanical pressure to ensure proper thermal conductivity if reused?
At 25:31 i see a bank of SMD-components that look like they 'code' the board version into the board. Either for visual check, or to be read by software.
A Zynq can't be nowhere near $200 in any significant volume, you can get a whole board including 1GB RAM and a 4GB SD card for close to half that
It dose have a FFT check out the data sheet, and they go up to 500 Meg
quality quality quality quality quality quality quality quality
Its Dave teardown detection they implemented with that heat sink detection :P
A snip at £ 4,560 plus 20% vat on top, I'll take half a dozen...
Whew, 350mhz 4 channel, triggered. Sweet! I want one.
Is another reason to use an LDO in a bench supply that has a switching pre regulator to have a small output capacitance?
You have a big cap on parallel with the load on a standard switcher.
Because they can!
Soooo when the EEVblog University will open in the new office? :))))))
Maybe they are using circuit designs from higher end equipment so that they don't have to pay to re-engineer it with cheaper parts. This isn't going to sell in huge quantities after all.
NI makes professional market gear with much higher specs than this. These are the circuits I was thinking of.
What are those metal boxes with the half open top? And what are they for?
Those 'blind' LEDs are probably 'secret until lit' - you don't see the legend when off.
+Graham Langley Bingo. Cleans up front panel . Better to hide things that aren't necessary. Similar to icons on your car dashboard. Out of the way when you don't need them, there when you do.
+Graham Langley Confirmed in the next video.
That's magnesium, not Al.
traps for young players have been catching older players from time to time ....
I think you like it
2 min into video and it already "paid off" for me. Extra earth screw on the back. I am currently doing a project that is gonna need exactly that. Why didn't I think of an extra earth terminal before? DOH Thanks!
+Chip Guy Vids They can be handy.
Hi Yah Dave. Love your tube though tell me? +/- 25V @ 1 Amp = how many Watts?
+260 NOB +25V*1А+|-25V|*1A=50W. Dave is absolutely right, as he is counting both negative and positive.
Yes cheers though am a little skeptical that the NI VB would produce this?
This sounds cool
BEAUTY !!!!
I teach a lab with these and trust me, students still find a way to kill them. Someone cracked a bnc connector off the main board, cost more to replace than I make all semester.