• About EEVblog

    Check Also

    EEVblog #1047 – Solar Roadways FINALLY BUSTED! Colas Wattway

    Dave finally puts an end to the idea of Solar Roadways. We have the test ...

    • Worf

      I can’t tell you how I know, but the P2000 platform is a custom Agilent internal reference design. They had two of them – one of them is the ARM based version like in the 2000/3000 series scopes with the ST Micro SPEAr 600 ARM processor, and the other is a PC-based one (hence the AMD chip). They’re supposed to be internal reference designs and customized to the final application, but I’m guessing it was just cheaper to use the already designed reference board than to build it into the main board like the ARM one is.

      I know the ARM based versions run Windows CE 6 or Linux, the PC version is supposed to run Linux I believe, or embedded XP.

      I’m sure the purpose of that PIC is probably as a USB keyboard controller – the front panel being the keyboard that sends USB HID reports to the PC host.

    • Rafael

      Sooo… That’s what a $20k teardown looks like, huh? Zillions of interesting things but still a few candies left for the imagination… 🙂
      I would have loved to see the BGAs under the fans, but a nice teardown indeed.
      BTW, the PTHs are very handy, especially for someone like me that completely forgot (and lost the damn book) of switching power supply design.

    • Sylvester

      I’m totally new to electronics. What do you use this device for?

    • AJ


      Great video, and wonderful to look inside such an beautifully designed bit of kit.

      As a novice when it comes high-end kit like that – I’m fascinated to know what makes it worth $20k to a professional user – what can it do that a $300 AFG can’t do? Is it just the expanded frequency range & accuracy / stability of output? Or is it more complex than that?

      Maybe if you do a “review” (or demo) video you could take a minute to two to talk about that?


      • It’s the performance. Bit rate, jitter, noise floor etc etc. High performance/high speed design stuff like say designing hard drives requires top end test gear. Stuff that only $50K-$200K gear can measure or generate.

      • jb

        Reading Agilent’s marketing material it’s clear that one part of the target market for this bit of test equipment is design verification of high speed serial data channels (e.g. SATA).

        For that purpose it can create a deliberately crappy signal (with known characteristics) that can be used to fine tune the design of line receiver circuitry to perform in worst case conditions. An interesting part of the specificaton is that the noise generator is not random but deterministic with a repeat period of about 20 days.

        A standard function generator makes a nice clean signal – something which is relatively easy and doesn’t need fancy kit like 2.5Gbit/s DACs and thee high speed circuitry needed to drive them.

    • stupid


      Could you provide pdf version for teardown,in China it’s difficult to view youtube.


      • Wartex

        Use TOR or buy a VPN node in USA.

        • Oliver

          Wartex of course ment the rest of the western world, not just USA.

    • Wow that looks really good! I am very curious about the operation of the unit. So Dave I look forward to the next video;-). GRATTEN also has a new Signal Generator is the GA1484
      If I have pictures of the inside, I shall share with you.


    • David

      Dave, If they let you keep it, I bet you’ll have to pay a big sack full of money to the Governement – right? Income and/or gift Tax?

    • f4eru

      Dave,take the rest appart, there is more good stuff just a few screws away ;)))

    • FCB

      Thanks for the brief teardown. I love bits of gear like this HOWEVER how does this fit in with your statement:

      “e.g. I’m sure your $50,000 power line analyser is a really great product, but it’s not an appropriate product for a general interest review.”


      • Because I have absolutely no interest in playing with a power line analyser or other obscure bit of kit. But a function gen or oscilloscope is right on the money, even if expensive.
        So it’s not about the cost so much, but about my personal interest and those generally of my viewers.

    • KTP

      Ah, those 7 pin TO220 size devices near the DC-DC boards are probably also DC-DC devices too. I haven’t bothered to look up their part number, but they look identical to the 12003 3 amp unit I am using on a proto board (love it, 93% efficient with built in inductor).

      • jb

        All the magnetics in the DC-DC converters are off-the-shelf parts. Given this is likely to be a low volume product it doesn’t make economic sense to use non-standard parts if a standard part will do the job. No doubt that’s why the reference design processor board was used.

        The Coiltronics VP4-0060 device is from the Versa-Pac range. These have 6 identical windings that can be configured in various ways.

        The Q4435-BL is a 1:1 transformer from Coilcraft from a range specifically designed for use with the LM258x flyback regulators – although it’s being used with the LT1370 here.

    • Salas

      The pricing of such products is down to their speciality of developement and the solutions they provide to a small target group of clients like defenece and healthcare electronics makers. BOM and execution costs do not look anything far north of a rack server computer IMHO.

    • jb

      The input power details vary according to the information source.

      The label on the back panel says 200-240VAC 50-60Hz and 100-127VAC 50-400Hz but the Agilent product datasheet says 100-240VAC 50-60Hz and 100-127VAC 50-400Hz.

      On the other hand the TDK-Lambda power supply datasheet says “90-264VAC (47-63Hz, 440Hz with reduced PFC)”.

      That can be accounted for by Agilent giving the nominal voltage/frequency values while TDK-Lambda show the maximum range for a mains supply with those nominal values – voltage ±10% and frequency ±3Hz (at power line frequencies).

      Likewise with the power rating there are multiple values. The back panel says 180W max, the Agilent datasheet says 90W nominal. The power supply is specified as 180W output (12V 15A). At its 90% maximum efficiency this implies 200W at the input.

    • Terry

      I have been looking at one of these and find it to be interesting, but it has contradictions in it. For example Triangle waveform is simply the 50% duty cycle of RAMP. There is no definition
      at least on line for the number of voltage points to make a Triangle/Sawtooth wave-form on the AWG side of things.

      I would like to see this instrument do the following, 70 Mhz Triangle wave-form, Two
      such wave-forms coupled on the same output together, with Amplitude Modulation on each individual wave-form, Phase Modulation, and Duty cycle/ pulse width modulation on each individual wave-form.
      How about it, hopefully you can post this diabolical test–lets see how it performs???
      Yes I do have a use for it, but Agilent are perplexed by this request when I ask them, can it do it???

    The EEVblog Store generally ships twice a week, on Tuesdays & Fridays, Sydney time. Dismiss