ADCMT 7480 or Advantest 6581T

[bastl_r]

I know it's a crazy thought.
But if I really wanted to immerse myself in the world of 8.5-digit measurement technology and was faced with the choice of buying an ADCMT 7480 or an Advantest 6581T, which one should I choose? The technology in both is roughly the same i think.
Which reasons would speak in favour of the ADCMT and which in favour of the Advantest?
Are there any reasons that speak completely against one or the other device?
What should I look out for in any case?
I don't really need an AC measurement.

Regards bastl_r

[Kleinstein]

The advantage of the old 6581 is that there is a reverse enginered schematics available, if needed for a repair. As a downside the displays are often dim and a replacement is a bit tricky.
A know weak point for the 6581 is the rather large current spike from AZ switching - this can be an issue with some signal sources. Another point is that ACAL may not be fully accurate for the highest resistor ranges. So the specs there may be a bit overly optimistic. Quite a few of the old meters also have a relay power saving mode not working right that can cause some thermal EMF error that can effect the lower voltage ranges - this is fixable at least.
The 6581T tends to have a rather poor factory calibration that causes quite some INL error - so while the ADC is essentially the same as with the full version, the linearity is not that great for the T version. There may be ways to correct this, but it is extra effort and may need extra instruments (e.g. cal source).

[branadic]

The 6581T tends to have a rather poor factory calibration that causes quite some INL error - so while the ADC is essentially the same as with the full version, the linearity is not that great for the T version. There may be ways to correct this, but it is extra effort and may need extra instruments (e.g. cal source).

I'm not sure this applies to the T models only, as I had the same issue on my D model. I would rather guess it's a general issue for all three models. The only thing known is that some nice INL figure was shown in some paper, where they probably applied individual INL constants to that specific meter, while all meters available on the secondary market seem to contain general and always the same coefficients.

-branadic-

[Mickle T.]

As I have already written several times, both of these multimeters are not of the metrological type, but are technological and are used in equipment for functional testing of microcircuits and semiconductors.
ADCMT 7480T (there is an error in the topic title - ADCMT 7480 does not exist) is designed as a cheaper replacement for the Advantest 6581T model. ADCMT 7480T:
- lacks current measurement,
- does not have higher resistor measurement ranges,
- does not support high-speed measurements,
- there is no documentation at all,
- circuit diagrams are still available only in the form of fragments,
- firmwares are not available,
- it is unknown whether there is a problem with INL and whether it is possible to solve it using software, as was the case with the R6581x.

[aronake]

Maybe you should not get too blinded by the 8.5 digits.

A Keithley DMM7510 may serve your needs very well. They start to come up used at not insanely crazy prices. Still a bit more expensive than an Advantest or ADCMT. But more linear, less noise, a modern user interface, documentation, spare parts avilable (as compared to the Advantest screen).

The DMM7510 also not a metrology multimeter. I think only 3458a, Transmille and the fluke 8.5 digits would classify as that. But it is not that far after 3458a in most measurement areas. Almost as low noise, better temperature stability, quicker warmup, a modern gui, smaller, lots of other features etc. It lack artefact calibration, but you likely will not adjust it yourself anyway.

Or just realize that you are falling into the voltnut abyss and save your self some future problem and buy a 3458a directly. If you find a good used one, you could see it as an inflation protected asset.

[bastl_r]

Thank you for your advice.
I thought I could delve deeper into the world of metrology for around €2000. For a DMM7500 I would have to invest twice as much. And a 3458a is significantly more than that.
The 6581T is actually already above the pain threshold.

Regards bastl_r

[dietert1]

A 3458A repair can be difficult to get and expensive (board level replacement). As a private person one may have to buy a spare 3458A. Many voltnuts have several of them, just in case.

Another idea:
Two years ago we got a Keithley 2700 plus some multiplexer plugins. It was € 450 and its voltmeter has excellent short term stability and good resolution when used with a GPIB or RS232 host. I used it for many interesting tests. The multiplexer can be used for nulling the meter (low thermal short on unused input channel). I showed some results in the thread about "multiplexer for voltage references". I also used it to test resistor arrays. It does 4-wire ohms over the multiplexer.
One should better spend some money on an incubator with a good TEC controller.
We also have a 7706 all-in-one plugin, that in addition to a multiplexer provides digital outputs and two DAC channels. Very useful, if you are able to make your own GPIB control software.

Regards, Dieter

[aronake]

A 3458A repair can be difficult to get and expensive (board level replacement). As a private person one may have to buy a spare 3458A. Many voltnuts have several of them, just in case.

Another idea:
Two years ago we got a Keithley 2700 plus some multiplexer plugins. It was € 450 and its voltmeter has excellent short term stability and good resolution when used with a GPIB or RS232 host. I used it for many interesting tests. The multiplexer can be used for nulling the meter (low thermal short on unused input channel). I showed some results in the thread about "multiplexer for voltage references". I also used it to test resistor arrays. It does 4-wire ohms over the multiplexer.
One should better spend some money on an incubator with a good TEC controller.
We also have a 7706 all-in-one plugin, that in addition to a multiplexer provides digital outputs and two DAC channels. Very useful, if you are able to make your own GPIB control software.

Regards, Dieter

Good suggestion!

Add a ltz1000 voltage reference on another unused channel, some python code with calculations and thats a metrology powerhouse!

[alm]

Note that there are other multimeters that can have a built-in multiplexer (like Keithley 2000/2001/2002, HPAK 34970A, HP 3457A, some Prema meters). And also stand alone switch mainframes like Keithley 7001. Limiting factor will be short term stability, tempco, linearity and noise of the DMM and uncertainty introduced by the switches (e.g. thermal EMF, isolation resistance, bias current, etc). You can find noise and linearity tests for some DMMs on this forum.

You can sometimes improve uncertainty by measuring differences (two voltage sources back to back) or measuring voltage instead of resistance (connect two resistors in series with a constant voltage across them, and measuring voltage across each resistor).