Author Topic: Advantest R6552 Multimeter Teardown  (Read 1833 times)

0 Members and 1 Guest are viewing this topic.

Offline chubbymonkTopic starter

  • Contributor
  • Posts: 28
  • Country: au
Advantest R6552 Multimeter Teardown
« on: January 24, 2023, 11:14:38 am »
First post! Hope the images embed properly.
Managed to get one in decent condition so here's a teardown. I think this is one of the later revisions(?), component date codes of ~2003.
Album with all the pictures and more: https://imgur.com/a/cWQYjfV

Front:
Nice and bright VFD display, the VFD really is very nice in person, has a bar graph as well if you're into that, membrane buttons feel meh, hard power switch and switch for front/rear input terminals.


Rear:
Rear input terminals, universal input voltage selector, Ext Trigger and Complete BNC terminals, GPIB and RS232


Two screws and the case and back slide right off. Revealing:


Nothing much to note of on the bottom, clear isolation between input and the digital side. Interesting looking FFC, first time I've seen a copper coloured one like that. Shielded maybe?


Taking off the two shields and unplugging the rather thin cables connecting the rear terminal reveals a proper look on the inside.


Taking off the extension board and unplugging the transformer cable reveals a good look at the main board. Taking the actual main board out seems like a real pain so no more disassembly.
Please correct me if I'm wrong, not 100% sure about everything. Closeups of various sections of the board are in the imgur album linked above.
Mains wiring is neat and tidy, good earth connection.
The metal plate still on the supports the input diode bridge protection, not much else in the means for input protection though.
U001 seems to be their custom ADC, heatsink is glued on.
U302 is a NEC CMOS-6A Gate Array and controls the ADC, function range control, interfacing through the opto-isolators to the digital side.
U107 is the voltage reference, an LT1027, specs don't seem to be terrific but I'm sure it does the job.
One big resistor network and a few smaller ones around the board, not sure of the manufacturer, did Advantest roll their own?
Xilinx CPLD for front panel keypad and VFD.
CPU is a Toshiba TMP68303F, hidden beneath the mains input (see pic in album) there's a row of DIP switches for some factory programming/testing I presume?
A sticker regarding a restricted rights legends for Mentor Graphics Corporation(Now Siemens EDA) which probably is the EDA which Advantest used? First time seeing such a sticker on an actual board.
Standard National Semi GPIB controller.


Extension board looks to primarily just be the AC/DC converter
Standard AD637 RMS to DC converter and your usual amount of analog switches.
Nice Panasonic caps (All electrolytics are either Panasonic or Nippon Chemicon, as you expect of Advantest)
There is a genuine bodge from factory though, a cap slapped between two pins of an Intersil component. It is nicely soldered though.
Isolation cutouts are quire rough though, almost like they're stamped out or something.


Some basic testing:
Only thing I have to compare is a fairly recently calibrated Yokogawa 7562, 6.5 digit class. I don't have any real precision voltage/current sources though so don't read too much into the readings.
If you assume that the 7562 is "correct" (not a great assumption) then it is a bit out, +0.007%-0.008% instead of the +-0.002% + 2-3 digits of the specs (24 hour DC voltage) but I'm sure the 7562 is a bit out as well and neither of them were truly warmed up properly.


Overall a pretty decent meter I think, especially when you can pick these up for absurdly cheap in Japan. Although international shipping to Australia certainly doesn't help any deal...
 

Offline Kleinstein

  • Super Contributor
  • ***
  • Posts: 15149
  • Country: de
Re: Advantest R6552 Multimeter Teardown
« Reply #1 on: January 24, 2023, 12:46:33 pm »
It would be unusual to have a slow ADC chip with a heat sink. The SD ADC chips usually have a rather moderate power consumption. A high power consumption is not working well with precision.
I would more expect some CPLD / FPGA maybe a custom chip to control a mulislope ADC of some kind - these sometimes run a bit hot.
It is still odd that there are some links to the long resistor array.

 

Offline Hydron

  • Super Contributor
  • ***
  • Posts: 1085
  • Country: gb
Re: Advantest R6552 Multimeter Teardown
« Reply #2 on: January 24, 2023, 12:55:28 pm »
My 34401A has a very similar factory bodge - leaded MLCC cap soldered between 2 legs of U101. Slightly surprising as both your meter and mine (2007) are late revisions.

Interested in how the ADC is done - agreed that it'd be surprising to see a low speed ADC with a heatsink.
 

Offline chubbymonkTopic starter

  • Contributor
  • Posts: 28
  • Country: au
Re: Advantest R6552 Multimeter Teardown
« Reply #3 on: January 24, 2023, 11:09:05 pm »
Yeah, on second thought it wouldn't make too much sense for U001 to be the ADC. Not sure which one is actually the ADC now...
 

Offline Kleinstein

  • Super Contributor
  • ***
  • Posts: 15149
  • Country: de
Re: Advantest R6552 Multimeter Teardown
« Reply #4 on: January 25, 2023, 10:50:50 am »
The the 6 digit level I would expect not to find a single ADC chip, but some multi-slope or similar ADC made from different chips, like OP-amps and switching chips (e.g. 74HC4053 in the fluke 8845 or HP34401) and some control logic in a µC or FPGA or custom chip. U001 may still be that custom chip and may even include some of the switching. There are plenty of precision OP-amps and a few hard to identfy chips.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf