Author Topic: Advantest R6145 Teardown  (Read 3173 times)

0 Members and 1 Guest are viewing this topic.

Offline free_electron

  • Super Contributor
  • ***
  • Posts: 7307
  • Country: us
    • SiliconValleyGarage
Advantest R6145 Teardown
« on: July 29, 2014, 06:39:49 am »
Got my hands on a used Advantest R6145 voltage/current source. These things are pretty unique. Not really usefull as a general purpose supply but very high quality and precision. The machine is essentially either voltage source with programmable current limit , or a programmable current source with a voltage compliance control. The output is fully push-pull and bipolar so in essence this thing works in all four quadrants.

built by advantest in Japan these sources are typically found in big ATE racks to test semiconductors. At 100$ in ebay this one was to good to let pass by.

Advantest is right up there with Agilent and Keithley when it comes to this kind of 'whacko performance' kit. Plenty of semiconductor manufacturing plants use large advantest macines to production test very complex mixed signal devices.
A brand to keep an eye on on ebay !

The machine was crusty on the outside. almost like it has sat under someone's turky fryer. coatedin sticky smelly cooking grease. A good wipedown with a mixture of soapy water followed by a polishing with hand sanitizer gel ( which is essentially 80% IPA mixed with a gelling agent and some glycol) restored the case and front panel to a clean state.
The guts themselves were sparkling clean. Not a speck of dust in sight. so this thing most likely spent its operational life in a class 10 cleanroom.

The beastie powered right up and still is spot-on. It offers several ranges. At the lowest voltage range (-300 to 300mV) you can adjust the output in steps of 10 microvolts ! At the highest range ( -60 to 60 volts) you still get to play in steps of 1 millivolt. Very respectable precision.

Current mode offers three ranges as well with the lowest range (-3 to 3 milliampere) offering a resolution of 100 nanoAmpere ... the highes setting (300 to -300 mA) is adjustable in steps of 100 microampere with max compliance voltage in the 60 volt range.
the machine can create , besides constant levels, burst pulses with two programmable levels, staircase or staircase bursts with a programmble amount of steps and levels. this supply has external sense lines as well as a guard terminal for when you are playing with precision stuff. The control loop time constant is programmalbe to dampen oveshoot.

This thing has some killer spec's ..

Output 1 Max. power: 10 Watts
Output 1/Range: 1 Max. Current 300 mA (300 mV to 60 V)
Load Regulation, Voltage: 0.0000024 mV    :o
Load Regulation, Current: 0.0000000012 mA   :o
Line Regulation, Voltage: 0.0000018 mV    :o
Line Regulation, Current: 0.00000003 mA    :o

All in all a very nice machine. i will finally be able to drive an LED without needing a series resistor  >:D

Picture time :

the frontpanel

the output terminals

300 milliamp range

3 milliamp range

60 volt range

300mV trange

The guts.

on the left front a classic power transofrmaer that makes +80 and - 80 volts DC post rectification and smoothing.
above that sits an HC 6303 cpu based (hitachi core) with a timer counter, TMS9914 for the GPIB a 32K eprom and a 8K static ram with 3.6 volt battery cell. The ram holds the calibration constants. Battery measured good. i need to take a ram dump through GPIB as i have no means to calibrate this kind of machine myself. I verified this machine against a calibrated 3458 and it is pretty much spot on.

The CPU section crosses a bundle of optocoulers into the control logic section.

The top board holds some control logic , a 12 bit multiplying DAC and a bunch of LM339 comparators

below that board sits the analog guts . A burr brown 16 bit dac as well as a dual 8 bit dac are the driving engines for the cotrol parameters.  most likely these set the outer ranges of the multiplying dac that then generates a fractional division of the settings, essentially building a weighted 20 it D/A converter. Note : speculation , i could not find schematics or manual for this machine

the analog section is peppered with marked parts. the LT1008 opamps have yellow paint dots on em. most likely these were hand-picked. planety of other components have matching colored dots as well. so there was a fair amount of hand-tuning this beastie.

the output section employs mechanical relays for range switching. there is no audible click when changing polarity of the output so the output stage is fully push-pull and 4 quadrant. the relays only click when changin operating range for voltage and current leading me to believe these relays change the gain of the end stage and the divider for the readback.

the output stage employs 4 beefy PNP and 4 NPN power transistors

all in all a very nice piece of test equipment that only needed a bit of spit and polish to make it presentable. works perfectly fine.
« Last Edit: July 29, 2014, 07:08:33 am by free_electron »
Professional Electron Wrangler.
Any comments, or points of view expressed, are my own and not endorsed , induced or compensated by my employer(s).

Offline EEVblog

  • Administrator
  • *****
  • Posts: 30635
  • Country: au
    • EEVblog
Re: Advantest R6145 Teardown
« Reply #1 on: July 29, 2014, 06:56:02 am »
Wow, $100!  :o

Offline Chalmml1

  • Supporter
  • ****
  • Posts: 5
Re: Advantest R6145 Teardown
« Reply #2 on: July 29, 2014, 12:05:52 pm »
A very nice score for $100 may it serve you well :-+

One thing to note about early Advantest gear is that almost all of it is based on previous designs under the Takeda Riken name before they rebranded. This machine is no exception but sadly a quick search still didn't turn up a manual, closest I could find was people selling a data sheet for the 6141 which seems to be the little brother of the 6145. No idea what if any useful info is in that sheet though.

Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo