EEVblog
Вставка
- Опубліковано 28 вер 2024
- Complete teardown of an ATL/Philips HDI 3000 ultrasound machine.
Testing video: • ATL HDI3000 Ultrasound...
Probe teardown: • EEVblog #1315 - Ultras...
Time Gain Control Compensation: www.ti.com/lit...
Subscribe on Library: lbry.tv/@eevbl...
Forum: www.eevblog.co...
#Ultrasound #Teardown
EEVblog Main Web Site: www.eevblog.com
The 2nd EEVblog Channel: / eevblog2
Support the EEVblog through Patreon!
/ eevblog
AliExpress Affiliate: s.click.aliexpr...
Buy anything through that link and Dave gets a commission at no cost to you.
Donate With Bitcoin & Other Crypto Currencies!
www.eevblog.co...
T-Shirts: teespring.com/s...
![[UA] Eternal Fire проти NAVI | EPL Season 20 Malta](http://i.ytimg.com/vi/chvlsDygrZM/mqdefault.jpg)
The BOM must look like a distributor's inventory stock. What is BOM management like at this scale?
thas why you have these dedicated boards that are already pretty complicated but not that many parts per board.
260 pins per plug, (yes i am that sad)
was gonna count but scrolled down first to check... thanks for saving me time ;)
@@tomaszkulawinski6388 Same here.
Yeah it's an ITT Cannon DL series connector, not "custom".
Chris W Any idea what the typical applications for such a connector would have been?
@@tookitogo We used them in patch bay connections for large broadcast audio mixing consoles.
Arg! As a sonar guy, I can tell you that the function of the TGC is to increase the receiver gain as time elapses from when the pulse was transmitted. Without this, the signal would be huge (and clipping) right after the transmit pulse ended, and buried in the A/D's quantizing noise at larger time delays from the end of the transmit pulse. In many radar and sonar units, this is called TVG (Time Varying Gain).
is basically equalization in non digital word I guess.
@@vaualbus I think "compression" is better word.
Gold-glad ceramic packages always feel special to me.
A real Bobby dazzler
I guess I'm kinda off topic but does anybody know a good place to watch new tv shows online ?
@Dangelo Terrance flixportal :P
@Byron Benicio Thank you, I went there and it seems to work =) Appreciate it!
@Dangelo Terrance happy to help xD
The vid at the begining, I could swear that I've seen it YEARS ago!
Thats because you did
Raytheon chips are hardened against radiation (RF, and Alpha, Beta, & X). They are used a lot in military applications.
They made the best vacuum tubes imo...though I doubt a lot of people watching this would care.
50 min teardown video? I like it!
Never apologize for the length of a video. Bored people can scrub around and see the bits they want to see. I appreciate the detail -- were it not for your teardowns, I'd probably never see most of this stuff (let alone get the insight!).
Fun fact, IDT was once an owner of the old AWA Semiconductor fab in Sydney.
Those Custom Connectors actually look like Cannon DL series. Used on some high end audio equipment (Solid State Logic 4000, 6000, and 9000 mixing consoles)
that control panel tho, mmm, trackball, nice looking keyboard, faders, giant buttons.. do want
Yes, ultrasonic cleaners and digital calipers are my favorite tools.
Those chips labeled "MHS" are mfg'd by "Multi-Health Systems Inc.", hence the acronym "MHS".
They are still in business and amazingly still making chips for ultrasound machines.
Torn ACL diagnosis... Ultrasound is not the right tool for the right job. Those sliders are the gain controls for different depths. I am totally enjoying this video!
I was wondering at the beginning 'damn, Dave is looking real fit and buff' :D
I used press fit connectors in the past, make sure you define the right hole size and they are super reliable. Order a couple of extra PCB's and you can press the connector yourself buy using a vise.
Pdsp16330a c0 Pythagorean data processor probably vector graphics angular converters.
Repairing this scanner and consequent dipping into it with oscilloscope, watching waveforms of the electrical signal that is being sent to transducer and received from the scanned media would be PRICELESS CONTENT.
Please, consider making such video before doing destructive teardown of probing heads.
I remember those connectors from ATE equipment, where they were used to connect the DUT cable assemblies to the front of the rack, and from there the connectors were split out to a set of power supplies, giving a variety of power supplies, 115VAC 400Hz 3 phase, each phase adjustable voltage wise, and with the actual output voltage read along with the current by both the power supply itself and the ATE acquisition system. Then 28VDC, up to 50A, as some of the stuff needed that amount of power to operate.
Then a massive patch panel, wheere you had a 200 by 200 pin array of interconnects, so you could have the input and output lines connected to the simulation and readout side.
Those connectors are rated for low noise, even after 1000 insertion cycles, and you really do not want to know the cost of them, though they do have the ability to replace individual pins and sockets if they do fail. Perhaps not with the ultrasound, as they likely potted the plug.
I would hazard half of those added modification wires are there to repair broken inner traces in the boards, as they likely were only tested and passed in the PCB plant, but after the PCB was assembled and soldered they would find the failed traces that did not survive the reflow cycle during test. Expensive board, expensive components but cheaper to do diagnosis and place the wires than scrap the entire board that mostly works, and hope the next board is perfect.
IBM Deathstar, not likely to still work, though there are a lot of glued down components, using some quality clear acetoxy free silicone instead of the crusty glue.
39:20 XWindow (no 's' at the end) is not an operating system. It's the GUI environment used on UNIX-like operating systems.
Finally you made this video
I can see why these devices used to cost so much. After the teardown, what do you do with all these boards? I can't see too many components that are usable in a modern environment.
When I was a kid, my dad fixed devices like this for Datex Ohmeda (senior service rep for BC and the Yukon). I don't know if he did ultrasounds, but there've been EKG, anaesthetic, ventilator machines, etc at our hose for repairs. Sometimes they'd ship to our house, sometimes he'd drive out to them.
I used to build similar machines back in the 80's under the brand names of UniRad or Johnson & Johnson. They rarely shielded the card cages like this, no need to if designed properly. They just put cans on the parts that might leak RF. Also, remember, these machines could take a black and white picture, derived from nothing more than sending out sound waves, and calculate weight at birth for babies. When the calculations were incorrect, the manufactures got sued for big bucks.
I always figured that the mathematical processing for radar might have a lot in common with that needed for medical imaging. I might be mistaken, but given that Raytheon makes radar equipment, that could explain why a Raytheon chip might show up in medical equipment.
I guess that the blown capacitor can be the issue of the machine. Maybye it short some power rail on the board. Coulkd you try to "reapir" it just to see if than the think work.
By the way I had watched that video the other day and think why Dave haven't do a teardown I guessed you throw away it.
Also I also saw the behind economic repair agilent DSA, that was a memories of the past.
There use to be ultrasound machines that had Apple G3 systems integrated into the cart. Very strange beasts to support. Running Mac OS with vendor hardware slapped on. Can't remember the company who made it. Think the Mac portion was more for handling the images. From outside you would never guess a Mac motherboard was hidden inside.
HP Sonos machines had those as an upgrade option. I believe the frontend was 68k based and used DSPs for video processing.
I'll be our friend over at Signal Path could explain how this works
This is so fucking cool
Now upgrade this badboy!
every product should incorporate design team engineers name. this is really a beast. at least we could know how many peoples were involved to get the job done. just like movies have name casting these kind of product really serving for mankind and the designers names should be published by the company.
Subject to the design engineer's consent. Not everyone likes to be named.
Those "optocouplers" at 39:48 are actually little 150MHz 1:1 transformers from Mini-Circuits
Wonder what would happen if you turned it on and used it without any shielding?
I was about to say "man i could have sworn i have seen this before"
Ultrasound is very operator skill dependent, I'm told. And e-net would be Ethernet, that's a nice old school thing nobody's seen since 10baset days. AUI connector or something like that?
Raytheon, missiles and ultrasound ?
Conflict of interests ?
Who made this beast device is awesome stuff
New equipment for bench pressing - great!
Wait, isn't EEVblog2 the work-out channel ...? Dave got me confused ...
Dave, as an engineer that in the past has messed with military stuff do you believe that this kind of electronics for medical applications has similarities with the military grade stuff in terms of design complexity and implementation quality? Thanks. Also, can you talk about military electronics and their piculiarities on an engineering perspective?
Would also like to see this. I'd bet there's a fair amount of overlap (both applications have life safety and reliability considerations). Guessing the military angle also has to consider the environment it's operating in? For example, a typical ultrasound machine can assume it's indoors at "room temperature." A typical military appliance won't be able to assume that.
According to the GNOME calculator's built-in unit converter, 200 kg = 441 lbs., so a good sized sumo wrestler.
How do you know how many layers the PCB is? is there a way to tell?
Has a lot more grunt than a QUANTEL PAINTBOX.
I had an ultrasound last year when I had trouble weeing lol
> this could be a 2h long video.
Well, we don't mind. And you could split that Video in 3 parts or so :)
The weird angle traces make me twitch a little.
Hello, your blog is very interesting. I am from Venezuela. I have a SIUI APOGGE 3500 ultrasound machine. I have not used the volumetric transducer for 4 years due to software problems. The software issues were fixed, but when I used the transducer again, the image it provided was broken (areas with images and others not). During the time that the transducer was not used it was protected, however, the place was subjected to high humidity. The other transducers are working properly and were used regularly. I would like to know where the transducer damage may be. Thank you.
So many busses, I was surprised it did not have a LNDN Bus! 😏
make diy harmonic scalpel
Heavy as a 7 foot wrester at 441 lbs.
It look like some server workstation
Woo I'm pretty sure I know one of the chips... the Symbios logic at 47:02 is a PCI / SCSI IO controller.
That Intel CPU looks like 486.
You also rent part of the basement?
25:05 There's the problem. No maybe but not probably. Is that a resistor pack? Could be sram.
i managed to get about 18 minuets in until my eyes glazed over and my brain shut down due to me not knowing what the hell you were talking about.
great vid though
lithium battery problem ?
That Intel processor looks like a 486 at time index 24:44
Can you confirm?
I have a whole box of them. It's 100% a 486.
@@An_Onion I thought later 386s came in the same form factor? Though, circa 1995, I'd also bet its a 486. With all that peripheral processing, anyone want to take bets on whether it's SX or DX?? 😁
The BOM on this must have been obscene back in the day.
43:20 forget about ordering your chips in a tray. just send them the board
I wish I had money to buy one
Today no doubt a single DSP chip and a driver chip do it all connected to a smartphone!!
what is a difference between a 12 volt supply (which has +ve and -ve terminals) and (-12)--0--(+12) volt supply (which has -12v, zero volt and +12v terminals). does it means, that there will be a 24v potential difference between -12v and +12v terminals? if yes then what is a 0v--12v--24v power supply? can you explain.
First learn about quadrants of operation of abstract voltage source. There are, obviously, four of them:
1. Positive voltage across the outputs, the unit is sourcing power.
2. Negative voltage across the output, the unit is sourcing power.
3. Positive voltage across the outputs, the unit is sinking power.
4. Negative voltage across the outputs, the unit is sinking power.
If the given voltage source (no matter, positive or negative voltage it supplies) can do both sourcing and sinking energy, there's really no technical difference between 0V-12V-24V and (-12V)-(0)-(+12V). The only difference is in output terminal labeling. That is only true for bi-quadrant/four-quadrant voltage sources, which is rare in practical: only high-end lab PSU, chemical voltage sources and capacitors are examples of such devices for DC, and only bare transformers - for AC.
Another aspect is _regulation_, i.e. an ability of the power sources to keep voltage accross some output terminals at the desired level.
In 0V-12V-24V case (let's assume terminals are labeled [A]-[B]-[C]) if you overload or short-circuit A-B pair, voltage A-B will drop to some extent, but the power source will keep trying to establish A-C at 24V. Voltage between B and C will increase, but nobody cares about Vbc.
In (-12V)-(0)-(+12V) case (same convention of term. labeling) if you overload or short-circuit A-B pair, voltage between A and B will drop, but the power source will at least keep trying to provide stable 12V between B and C. Voltage between A and C will, of course, go lower than initial 24V.
So, it depends. The way terminals are labeled usually remarks the way their voltages are regulated (usually 0V is a reference point for all other voltages, but not always) and energy flow direction (usually positive terminals can only source current and negative terminals can only sink current). The GND/0V terminal in bipolar/multi-polar voltage sources can both source and sink current, which is the main reason why dumb relabelling of terminals can't convert 0V-12V-24V into (-12V)-(0)-(+12V).
@@toxanbi Thanks for the detailed explanation.
Ps somebody could create a firmware for this system and amalgamate all this graphics processing power into one heck of a gaming system..
Don't drink and route
Really missed this type of long teardown video
shot in 4k yet video is in 1080 wtf
The real question is why the heck did Dave buy this beast? Sure, he couldn't find one in the dumpster room, but . . . it was nonworking at the time, probably prohibitively time-consuming and/or expensive to fix. Must've cost a fortune to ship. Hm . . .
E-Net - must be one of those newfangled Ether Net connections - so etherial, I'm almost flying!
AUI Ftw :)
24:57 heatsink? hhaha this is some strange heatsink also werent the old cpus low clock speed enough to just dont generate that much heat afterall?
15:34 all the clips are not undone i see one at lest not undone
Hey, this isn't the quantum computing video.
9:45 that's what she said
"What is a 'ute'?"
-My Cousin Vinnie
Essentially the Australian equivalent of a Pickup Truck (more or less)
Utility car/truck
Came here for this. Was not disappointed.
6:25 E-net is Ethernet: that’s an AUI connector, to which one attaches an Ethernet transceiver for the physical medium used (10Base-5, 10Base-2, or 10Base-T).
Been a while since those were popular. I do retro-computing for a hobby, and I don't have too many cards with AUI connectors.
Please tell me your sarcastic.... in the early 90’s everything had AUI.. I made lots of money (for my bosses) selling AUI transceivers to swap users from 10Base2 to 10baseT
I'm not the only one, although I think I'm the only one who decided to use 75 ohm coax to set up a test Network...
Came here to say this. en.wikipedia.org/wiki/Attachment_Unit_Interface
@@imark7777777 Supposed to be 50-ohm for both 10-base-5 and 10-base-2. CATV coax is 75-ohm.
I guess the E-NET is Ethernet but 10Base5 (AUI), link:
en.wikipedia.org/wiki/Attachment_Unit_Interface
Correct, it’s AUI. It was used for 10Base-5, -2, and -T with the appropriate transceiver.
Bingo, I still have one of the adapters.
briefly played with some old ethernet cards with this connector and the coax 10Base2 also had some adaptors for AUI to 10base2 coax. I actually set up a small 10Base2 network before I had the gear to make 100baseT etc UTP cables. Played with an old print server that only had a 10Base2 interface, had to talk to it with telnet at one point... good times
@@johndaugherty7465 I do too
william fleete 10Base2 was also my first exposure to Ethernet. But it was already on its very last legs at that point, with twisted pair clearly having won the battle! :)
Assuming it's standard SCSI, you really should rip out and hang on to that MO drive and disk - those are small and rare enough to be interesting to a computing museum.
44:00 on the Advnaced IF Output board top left corner the two ceramic chips are Cartesian data (Real and Imaginary) into Polar form (Magnitude and Phase) converts (PDSP16330A) made by GEC-Plessey Semiconductor (GPS).
You can easily buy an ultrasound that has more capability than this which will also fit in a briefcase and weigh about as much as a heavy laptop. Ultrasound technology has come on leaps and bounds in recent years. For an example look up the Philips CX50.
I saw an ad prior to this video for a probe that plugs into a phone, WTF!
Now I want to see a teardown of that thing, but I guess I have to wait another 25 years or so for that ;)
My attending has a ultrasound probe that plugs into his iPhone. Yes, my friends, Dr. McCoy's Medical Tricorder is a reality...
Whoa. So all of the signal processing done by those shielded cards fits in a laptop size enclosure? That's awesome
Believe me that that hdi 3000 has a doppler spectrum quality competing with todays mid-high level ultrasounds (not high ends, but very close). If it had some B mode image processing capabilities of CX50, iHDI3000 should have been better than it in B mode too. Because it has 128 channels using all trnacducer crystalls without switching and a very good analog signal amplifier. Channel number is more than most of todays machines, so this rockets the signal quality. . Yes it is some old. But for the age of it it was like from outer space. I have used one 10 years ago, and shocked about its capabilities. First of all probe quality portable machnies are poor. They generally use a new tecnology (without crystal), I am not sure about cx50 but HDI porobe quality is top one, single piece crystal cut. If the pure signal quality recieved is good (signal/noise ratio), it really boosts image quality. Todays lower end ultrasounds mostly use some mathematical calculations like speckle noise reduction and some other techniques to boost image quality. I have worked on ultrasound (as field service eng. and application specialist (Medison- Samsung, Kretz, Hitachi-Aloka) machines since 1996, and saw lots of USG machines.
10:00 Dallas1286?
Thats why they threw it away! After decades the Battery is empty!
Helped another Doc save his unit by changing that stuff !
He didn´t want to learn how to handle a new unit because of his age...
That's what I immediately thought when I saw it. Cause of failure - memory battery.
Wasn't there a battery next to the module on the CPU board?
@@ve2mrxB These dallas modules are notorious, the battery is encapsulated inside the giant package. It wouldn't be using an external battery even if there was one.
8:15 "so many boards and so many things we can talk about this could be 2 hour long video"
Why not?
The chassis looks like it's international space station ready. Very well built machine.
Yep pretty confident the space station was very intentional
I had the pleasure of buying some medical equipment caster wheels as surplus. Strong as hell and silky smooth, the very best.
This might actually exceed the capacity by weight of many orbital launch platforms.. Are you thinking like space anchor or something ?
Am I watching mikeselectricstuff?
No :)
You are watching EEVblog.
Who?
@@woldemunster9244 mate, you are joking aren't you? :)
No, you can tell by all the tired catchphrases.
hah googling this thing results in EEVBlog links being the top hits :)
Wife blew out her ankle at netball.
Went to Doctor “ how many days” not our GP as it’s just a normal netball thing, no big deal..............
X ray. No busted bones.
Ultrasound. All good no problemos.
Still not better after ages.
Go to our GP and he gets a bit pissed and orders an MRI, ligament no longer connected.
One full ankle reco later.
Ultrasound was piss poor for this scenario
Bruh.
Ultrasonography requires an exquisite understanding of anatomy and physics and the ability to interpret three dimensional imaging. It takes years of practice. It's all about the gorilla holding the hammer, as AvE would say.
Yeah it's a bit of an artform, I find it's best whenever you break something you break it well enough that it's obvious to even the first year med students... When they can point to two different parts of the picture at least 5cm apart and say this is where it is and this is where it's supposed to be then it gets diagnosed quickly
Sonographers are not doctors, it generally requires a masters degree in ultrasound. Due to this, sonographers, particularly new ones can miss things, especially ligament damage which can be hard to spot. In this case, the GP should have either been able to diagnose a torn ligament or asked for an MRI straight away. More than likely a bad combo of bad doctor and bad sonographer, but can't entirely blame the sonographer.
@@ShortCircuitGaming a sonographer should not be trying to diagnose this kind of thing. Nor should a GP. This is not a simple pregnant/non pregnant diagnosis. This kind of diagnostics is the domain of the specialist radiologist. This may sound overdramatic but with ultrasound you need to "become the machine" and have scanned hundreds(thousands) of normal before you can identify abnormal. So many variables at play.
I was having elbow pain, they found it on sonography and knew it was a tendon issue.
SYMBIOS LOGIC? That looks like a SCSI Host Controller. A dead Givaway is the SCSI LED a thumb or two next to it ;)
Yup, SCSI bus controller. They were sold under a number of names, Symbios logic, LSI logic and NCR (I think they were the original owners of the IP). The 53c710 is a really old one, apparently also found in amigas but they were everywhere. I remerber wanting a 53c896 based card for my alpha.
The HDD is an LVD SCSI-3 drive, so it's probably for that.
Symbios logic 53C710 is a SCSI controller. I have a 53C865
At 45:40 you can see an x86 Processor by AMD.
An 80186-16. You don't see them all that much (or rather more outside of an IBM compatible PC than inside one)
"What's e-net?" Pretty early ethernet connection. Looks like a pretty standard vme chassis and backplane for high frequency and low noise operation and computing.
Correct. It's actually an AUI connector for 10Mbit/s ethernet. You'd need to have used a transceiver to convert it from AUI to BNC or RJ45 dependent on how your ethernet cabling was wired at the time. FWIW, the original Cisco Routers (2500 series) used to have only AUI on them for years.
WOW! That machine is absolutely amazing! My mind is blown. SO many ICs. Its never ending. Think of the people that designed and layed out these boards. Absolute geniuses.
Remember, this was designed in the MS-DOS era, there was no GUI CAD software 30 years ago. It must've taken months if not years to get it right. The schematic diagrams must be 1000 pages long.
He should frame some of those PCBs and put them on the wall. I would.
Simon. There were GUI CAD systems around more than 30 years ago. I was designing PCBs on a very sophisticated Computervision CADDS3 system back in 1984 and even then that system was already pretty old. There were plenty of other CAD systems around at the time although these were running on mini-computers or even main-frame computers.
This video is a long time in the making. I've been waiting for this video for ages thankyou!
6:26 That E-Net is using an older style Ethernet connection using DB series connectors. We used them on some of our network equipment back in the 1990s.
I used to service these units when they were still in use. They weren't as popular as the higher-end HDI 5000, as they didn't have as good of image quality. ATL was purchased by Philips Medical, BTW, so they still exist within that corporation. The panel on the front was referred to as the "mux" panel (multiplexer). It is responsible for switching between the three transducer input connectors. Those highly shielded boards the the mux panel plugs into are called the front end boards. They are responsible for beam forming. Each transducer is a bit different (i.e. linear, convex, sector, etc.) and these boards compensate for each individual channel on any given transducer. Each front end board contains an array of preamps that are controlled by the software. They can be swapped around in order to troubleshoot faults in the boards. Information for each individual transducer is stored on a circuit board inside the large connector for that transducer. That's why the connector housing is so large. Color, PW and other functions are added by the boards connected to these front end boards (just keeping it simple). By the time the signal leaves the front end section, it has been processed into a raw image with X,Y,Z and direction of flow information. This then goes to the "back end" section of the unit, where the image is processed, demographics are added and the image is displayed on the monitor. The MO disk on the back of the unit is sometimes a bootable disk. It can be used to restore the hard drive when it gets corrupt. I can't remember all of the commands, but I think one of them is when you press "super key" and "0" at the same time, it will put you into service mode. Thanks for the tear-down video. It brought back a lot of memories!
@45:41 Did you just call an AMD processor an intel processor? Oh, the blasphemy!
It is a licenced (has a "(C) Intel" on it) - 'cause second source and such - 80186 by AMD. ... While manufactured by AMD it was designed by Intel. Thus, cut him some slack ... :)
For the OCD:
The N80C is: "redesigned CMOS version that operated at 16 MHz and introduced a number of enhancements in addition to being lower-power CMOS, including a DRAM Refresh Control Unit and various power saving modes" ... but AMD and Intel sold the same product: 'N80C186-16'
www.cpu-world.com/CPUs/80186/Intel-N80C186-16.html
en.wikichip.org/wiki/amd/am186/n80c186-16
en.wikichip.org/wiki/amd/am186
www.cpu-galerie.de/html/intel80c186.html
What I really wonder is if the ceramic Intel under the heatsink is actually a i860 or i960?
@@darknase The one on the other board, with the round heatsink? Yeah, I'd like to know that, too! Old Intel CPUs (386, 486) also looked just like this.
I thought I was interested in what looked like Tracy Island in the first part of the video, but then I saw the Sinclair C5 in the background at 2:14!
Richard Garnish the C5 is the reason he's cleaning up his storage.
Dave, fix up the C5 and chuck out the old buggies!
@@Mark.R_ He's actually done a video on the C5, several years back.
Wow, this is a blast from the past for me! Was working on these when I got into ultrasound field service. If you want to see newer examples, we've got a ton of them around here I have access to.
AMI is American Megatrends Inc, the people who used to write BIOSs amongst many other things.
More likely it is American Microsystems Inc, a manufacturer of MOS-based ICs
E-net is actually Ethernet. That's an old AUI transceiver port. Plug in your 10 base-t transceiver and you'll be right as rain.
Xwindows is not an operating system.
The PDSP16630 is a very interesting chip...Plessey called it a Pythagoras Processor. It does conversion of Cartesian coordinates to polar coordinates.
The control panel wants repurposing as a Kerbal Space Program control panel! =)
Dave. Have you done a tear down of the HP9845 calculator appearing in the background at 2:15. Could be an interesting device to have a look inside.
I seem to remember HP devoted a whole issue of the HP Journal to describing the design and construction of this calculator when it was released.
Was just going to ask if that was an HP 9845 with the monitor missing.
Not a calculator though, a serious CAD workstation - for its time. hp9845.net/9845/hardware/9845c/
@@pixymisa8087 I called it a calculator as it is featured in the calculator section of the HP catalog. However, with closer reading of the catalog, I see that HP describe it as a desktop computer.
The traces on those power boards look really unprofessional. Maybe they had an intern who could lay those boards out while the pro's worked on the real stuff? The rest of the system looks great.