Advantest R6581T 8.5-digit DVM/Ohmmeter repair



Work in progress, this article may contain errors and mistakes

This article is about japanese-made Advantest R6581T review, repair, calibration, service and performance tests. We slowly getting to complete coverage on all 8½-digit DMM/DVMs that were in production, with this R6581T worklog article. Here’s a complete list 8½-digit long-scale meters that we know about:

This is 19” rack-mount size system DMM, just a little smaller than competition Keysight 3458A or Datron 1281. It is however much lighter, due to use of aluminum chassis and shield sheets, instead of heavy-duty steel frame in HP meter.

Manufacturer Model Type Status
Advantest/ADCMT R6581 (T/D) DMM No longer available for sale
ADCMT 7480T DVM,Ohms Current model
Datron 1281/1271 DMM No longer available for sale, obsoleted by Fluke
HP 3458A (+Fluke HFL version) DMM Current model
Fluke 8508A DMM Current model
Keithley 2002 DMM Current model
PREMA 6048 DMM No longer available for sale
Solartron 7081 DMM Also known as Guildline 9578. No longer available for sale
Transmille 8080 DMM Current model
Transmille 8081 DMM/Electrometer Current model
Transmille 8104 DMM/Electrometer Current model
Transmille 8881 DMM Current model

R6581 is especially interesting, because it’s design was direct response to mighty HP 3458A, and lot of ideas and features are directly featured against the HP meter. In some parts this meter is better, while in others worse. Take a look on specification. Specification from HP 3458A (002 option) is shown in blue color.
There are three main versions of the Advantest R6581:

  • R6581 is metrology type instrument with full DMM functionality, for DCV, DCI, 2/4-wire resistance, TrueRMS ACV and ACI.
  • R6581D is field type instrument with DC only DMM functionality, for only DCV, DCI, 2/4-wire resistance, without AC module.
  • R6581T is field type instrument with DC only DMM functionality, for only DCV, DCI, 2/4-wire resistance, without AC module. team got hands on R6581T version of the unit, which is equivalent to R6581, but without AC TrueRMS option board.

Function Range
DC Voltage 100mV 1 V 10 V 100V 1000 V
Accuracy (1y), Rdg ppm + Range ppm (HP3458A/002) 7+10 (5+3) 6+1 (4+0.3) 5+0.2 (4+0.05) 7+1 (6+0.3) 7+0.2 (6+0.1)
Temperature coefficient, ppm 1.2+1 1+0.1 0.5+0.01 1.2+0.1 1.2+0.01
RMS noise at NPLC100, Range ppm 0.1 0.05 0.05 0.05 0.05
Input resistance, Ω >1010 >1010 >1010 106 ±1% 106 ±1%
AC Voltage (Model R6581 only) 10mV 100mV 1V 10 V 100V 1000 V
Best accuracy (1y), Rdg ppm + Range ppm (HP3458A) 300+2500 (200+110) 200+200 (70+20) 150+1500 (200+20) 500+200 (400+20)
Temperature coefficient, ppm 10% spec\°C 10% spec\°C
Bandwidth frequency (max accuracy) 1 Hz to 2MHz (100Hz to 10kHz)
Resistance 4W 10 Ω 100 Ω 1 KΩ 10 KΩ 100 KΩ 1 MΩ 10 MΩ 100 MΩ 1 GΩ
Source current (High power mode) 10 mA 1 mA 100 µA 10 µA 1 µA 100 nA 10 nA
Accuracy (1y), Rdg ppm + Range ppm (HP3458A) 14+20 (15+5) 12+5 (12+5) 10+5 (10+0.5) 8+0.5 (10+0.5) 14+1 (15+2) 50+5 (50+10) 500+10 (500+10) 5000+10 (5000+10)
Source current (Low power mode) 10 mA 1 mA 100 µA 10 µA 1 µA 100 nA 10 nA 10 nA
Accuracy (1y), Rdg ppm + Range ppm 14+20 14+20 10+5 12+5 14+5 50+5 500+50 5000+50 5000+10
Temperature coefficient, ppm 3+3 3+1 2+1 2+0.1 20+0.1 100+1 1000+1
RMS noise at NPLC100, Range ppm 0.1
DC Current 100 nA 1 µA 10 µA 100 µA 1 mA 10 mA 100 mA 1 A
Accuracy (1y), Rdg ppm + Range ppm (HP3458A) 25+400 (30+400) 20+40 (20+40) 20+20 (20+10) 20+20 (20+8) 20+20 (20+5) 20+20 (20+5) 30+20 (35+5) 110+20 (110+10)
Temperature coefficient, ppm (HP3458A) 10+200 (10+200) 10+20 (2+20) 2+4 (10+4) 5+3 (10+3) 5+2 (10+2) 10+2 (10+2) 20+2 (25+2) 20+3 (25+3)
Shunt resistance (burden voltage) 1010 kΩ(55mV) 105 kΩ(45mV) 10.1 kΩ(55mV) 1.01 kΩ(75mV) 102 Ω(100mV) 12 Ω(100mV) 3 Ω(250mV) 2 Ω (<1.5V)
RMS noise at NPLC100, Range ppm 2
AC Current (Model R6581 only) 100 µA 1 mA 10 mA 100 mA 1 A
Best accuracy (1y), Rdg ppm + Range ppm 200+200 800+200


Redistribution and use of this article, any part of it or/and any images or files referenced in it, in source and binary forms, with or without modification, are permitted provided that the following conditions are met:

  • Redistributions of article must retain the above copyright notice, this list of conditions, link to this page (/fix/r6581t/) and the following disclaimer.
  • Redistributions of files in binary form must reproduce the above copyright notice, this list of conditions, link to this page (/fix/r6581t/), and the following disclaimer in the documentation and/or other materials provided with the distribution, for example Readme file.

All information posted here is hosted just for education purposes and provided AS IS. In no event shall the author, site, or any other 3rd party, including Advantest be liable for any special, direct, indirect, or consequential damages or any damages whatsoever resulting from loss of use, data or profits, whether in an action of contract, negligence or other tortuous action, arising out of or in connection with the use or performance of information published here.

If you willing to contribute or add your experience regarding Advantest/ADCMT instruments repairs or provide extra information, you can do so following these simple instructions

Manuals and documentation

The only available official manual data is in japanese language, but thanks to modern OCR and online translators, we can get most of important information about the meter.

There is no service manual or schematics available for this unit.

However great deal of useful information can be found in this thread at forums

Exterior and user interface

Teardown and design evaluation

Digital and interface PCBA

Analog PCBA

Master DC Voltage reference PCBA

Inguard power supply PCBA

Front panel PCBA


Troubleshooting and repair

h2 Firmware backups

TL866CS programmer with PLCC socket adapter was used to read back all firmware ROMs. Provided for use AS IS.

Firmware dumps received from internets:

Direct calibration vs HP 3458A and xDevs standards

External calibration for the Advantest R6581T was performed with help of xDevs’s primary standards:

  • +10.00000 DC Voltage standard, 792X prototype build around 792A Power Pack and our own LTZ1000A-based buffered reference.
  • 9999.9747 Ω ±0.33 ppm Fluke SL935, which was calibrated November 28, 2017.

Same standards were used to calibrate Fluke 5700A which was used as reference source for all ranges and functions verifications. Calibrations of both MFC and meter were performed on same day, so long-term drift can be ignored and only relative accuracy specifications are used.

Calibration accuracy verification

This section covering non-factory calibration procedure and testing of 24-hour stability performance on the meter. This is non-traceable calibration for education purpose only. By no means using even high-end calibrator can provide legally valid calibration certification.

To run some performance numbers on Fluke 5700A, let’s connect one to Fluke 5700A working standard multi-function calibrator and use’s calkit Python application to crunch some numbers.

Image 66: Advantest R6581T 8½-digit DMM with Fluke 5700A MFC during the calibration.

Reference source value = (Low limit + Hi limit) / 2;

DCV Function performance test

Here is automated performance verification test data before the calibration adjustment.

Before calibration adjustment, AS RECEIVED condition. Last calibration date or history of the unit is unknown.

DCV Test R6581T Data Source uncertainty Low Limit Hi limit Measured 24hr spec Result
0.1 VDC (0.1 Range)

Table 5: R6581T fresh after external calibration

As we can see, meter cannot meet it’s 24 hour specification for 3V and 30V ranges, but actually not that much off. 300 mVDC range was not tested due to bug in control code.

Resistance function test (4-wire with offset compensation enabled)

OHM Test R6581T Data Source uncertainty Low Limit Hi limit Measured 24hr spec Result
1 Ω

Table 8: R6581T performance test, as returned after calibration adjustment, 4-wire resistance

Test on highest range with 1 GΩ standard was performed using external standard, as 5700A cannot source 1 GΩ.

Calibration test for another unit

EEVBlog member Ramon contacted me regarding improving uncertainty and measurement ability for his DIY resistance wire tests and projects. He bought another R6581T locally, and we had short meetup to perform calibration accuracy check and get some data comparison going.

792X prototype 10V reference as DCV source, Keithley 2002 #2 as reference meter were used to do checks. Together with reverse verification to my primary HPAK3458A meters, this setup allow to verify performance of R6581 with relatively good uncertainty.

Brief teardown of his Advantest was also performed, with photographs presented below.

Transformers are all good and happy, overall unit is very clean and tidy.

Firmware version of this unit is very same to our own, so there was no action taken to back it up.

Unlike xDevs’s unit, in this one battery definately does not like factory-made installation, so it was replaced at some point. It still has juice and measured over 3V, so no problem here, for now.

Restoration summary

This is quite interesting instrument with it’s own quirks and features.
Definately a grab if you see one at low price under few hundred USD.
Here is small table of featureset against Keysight 3458A and Keithley 2002:

Feature Advantest 6581T HP 3458A Keithley 2002
Performance Good Excellent Good to excellent
Input terminal type Non-standard banana jacks 5-way binding posts Banana jacks
ADC DCV Linearity Okay Best on the market Good
Ratio measurement modes Yes, front/rear Yes, force/sense inputs No
Guard input Yes, floating or tied to LO No
Artefact calibration Full, with 4W Zero, 10VDC and 10 kΩ only AC only, need MFC fo DC/Resistance
Front/rear switching Electronic, using relays. Front panel/GPIB Mechanical switch only
Internal temp sensor Yes
Display Dual-line VFD Single-line VFD 16×1 Dual-line VFD
I/O GPIB only, SCPI GPIB only, custom protocol GPIB only, SCPI
Expansion options Analog output, Scanner cards None Scanner cards, nanovolt Keithley 1801 preamp kit
Cooling Active cooling with fan and airducts
Form-factor, weight 2U 19”, light 2U 19” heavy half-19” rack, light
In production? No Yes

Also some generic parts were ordered from DigiKey store, such as DC fan, capacitors. Here’s order list (there were some other parts for other projects, not shown here, to make up for >100$ bill, as then it’s eligible for free shipping to Taiwan :).

Quantity Part Number Description Extended Price USD
1 565-1846-ND CAP ALUM 680UF 20% 50V RADIAL 2.46

Table 22: First Digikey order for parts

Total repair cost in parts as for today: $1312.64 USD

Date Activity Time spent
02/26/2018 Received unit, initial teardown 1 hour
02/28/2018 Repair evaluation, testing Outguard boards and PSU, fix VFD panel operation 3 hours
03/02/2018 Ordered caps and from DigiKey 1 hour
03/07/2018 Received parts, replace all electrolytic caps 2 hours
Taking photos, writing up article sections and posts 5 hours

Table 23: Time worklog summary

Total man-hours spent on this project around 24 hours, give or take few. Was it worth doing? That’s up to you, so please leave your comments, so I can do more stuff like this.

Equipment used during restoration project:

I would like to express my big respect to “Mickle.T” EEVBlog member for his time and efforts on reverse engineering R6581 design in great detail. This article would not be possible without you. Спасибо за проделанную работу!

Discussion about this article and related stuff is welcome in comment section or at our own IRC chat server: (standard port 6667, channel: Web-interface for access mirrored on this page.

Author: Ilya Tsemenko
Created: Feb. 28, 2018, noon
Modified: Jan. 28, 2020, 11:51 p.m.