Advantest R6581 8.5 digit DMM mini teardown/repair

[little.tesla]

Thanks for the prompt replay!
I'v already reviewed those pictures, but there are no details how this copper resistor is constructed and of what resistance it should be.
I understand that R234 in combination with the OpAmp is used to boost the Ref voltage and any variation in the ratio changes the Ref voltage.
Do I aim to adjust the ratio to optimal 7K32/10K or to a specific Ref voltage? Or does the added copper resistor compensate the TC, but then what value should it have?

[branadic]

Now I got your question.
The resistance you aim for varies from meter to meter and its t.c.. So there is no general fixed resistor value to aim for, but individual measurement and adjustment. Instead you need a second multimeter and a large thermal chamber. Put your R6581 into the thermal chamber and run a temperature sweep. Measure the ref voltage of the R6581 with the second multimeter.

Now it depends 1. whether your ref voltage has positive or negative t.c. to answer the question to which leg to add the "copper resistor" and 2. what the absolut value is to calculate the required copper resistance and eventually the diameter plus the length of the wire needed. There is this tool I link to below that might help you to understand the temperature compensation.

https://www.eevblog.com/forum/metrology/adr1399-reference/?action=dlattach;attach=1467001

-branadic-

[Kleinstein]

One does not really need a 2nd meter for test. I would even prefer an external stable reference (7 or 10 V) and use the 6581T to measure that fixed reference. So the TC adjustment for the whole meter including the ADC gain and not just for the amplified reference.

[branadic]

That is a valid possibility, but has one disadvantage that is, while you could compensate for voltage readings in the 10 V range this way things could get worse for resistance or current readings.

-branadic-

[Kleinstein]

That is a valid possibility, but has one disadvantage that is, while you could compensate for voltage readings in the 10 V range this way things could get worse for resistance or current readings.
-branadic-

That's true, but usually the resistance and current ranges are not that stable anyway.
The restance and current ranges still contain the ref. and ADC part, just additional components (gain and the current source or shunts).

A first point is checking how bad the TC actually is.  There is still a chance that it is not worth correcting an only small TC (e.g. if < 0.5 ppm/K).  The temperature effect can also be nonlinear and the added copper can only correct for a linear part.

[dietert1]

Recently i have been using a Keithley 2700 with a 7706 scanner to characterize UPW25 precision resistors. Those resistors are specified for TC < 5 ppm/K. I found that the meter contributes about -2 ppm/K for resistance measurement. So its internal reference resistors are pretty good. Yet the meter exhibits a complex reaction to temperature change: Not only a linear TC but also some differential contribution. I mean a deviation proportional to the speed of temperature change. A typical time scale is two hours until these additional deviations vanish.

So this means either the resistors to test or the meter has to be inside a thermal chamber. The K2700 fits into a HN-2 incubator, so that's an easy one. And with those 77xx plugins it's easy to monitor the internal temperature. I won't try to mod the meter, but understand the deviations and subtract them in data analysis.

Regards, Dieter

[little.tesla]

I understand the reasoning and recommended testing approach to tune the TC of a specific meter. Thanks for the pointers.
Maybe I'm missing something, but I'm wondering what was the initial approach from Advantest. If I recall correctly the different models R6581, R6581D and R6581T are HW wise almost the same (apart from the small differences previously discussed) and it's assumed that for the higher grade models the critical components have been selected with tighter tolerance and maybe even lower TC. But this would mean that no specific TC tuning was done on the HW. I would assume that it's less likely that there is a software TC tuning of a specific unit. Probably Mickle T. would have found it.

I have currently two R6581T meters and when I did a test to monitor the same 10V reference, both meters drifted in opposite direction for about less than 1ppm within a few hours time. The temperature in the room is quite stable within ~1°C per day variation and I've measured the 10V reference with two calibrated 3458A (one with opt02) and know that the short term variation over ~10h of time is less than <0.1ppm. Over the weekend I added a driver for the meter to my MTE automation and I'll collect some data and try to share it when I have some more time to have something more scientific.

[dietert1]

From what you wrote one can understand that the voltage measurements of your R6581Ts may have a TC of 1 or 2 ppm/K - a good result. And i doubt you can derive a 0.1 ppm observation using HP3458A meters. If i remember correctly they spec about 0.6 ppm in 24 hours or so - and certainly with requirements concerning temperature and humidity.
A 8.5 meter is made from similar parts like 7.5 or 6.5 meters so it has similar uncertainties. To do metrology at the 1 ppm level or below requires a good meter but also equipment and parts not present in a DVM.

Regards, Dieter

[Mickle T.]

But this would mean that no specific TC tuning was done on the HW. I would assume that it's less likely that there is a software TC tuning of a specific unit. Probably Mickle T. would have found it.

I have reversed the entire R6581 firmware and can say that there is no TC software tuning in this multimeter

[little.tesla]

Your comments are greatly appreciated!
I understand that I'm moving on very thin ice and I don't have enough data to claim something, nor do I own the necessary equipment for doing this. At the moment I'm a bit exploring in which direction I should/want to go and was wondering about the TC stability of the R6851T meters I have and possible improvement. A bit to explore the borders of my R6581T meters.

I've attached below plots of my two meters sampling the same 10V source with 100NPLC over night. In the x axis we have samples. The test setup on my bench is shown in another picture. The sampling was started after about ~5h warm up and an internal calibration of the meters. The temperature in the room is stable to within +/-1°C, I did not have time this morning to download the data of the temp/humidity logger. Maybe the captured data is useful for others. I started this morning another acquisition, let's see how stable it's over the day.

I do not want to go offtopic, but for reference I've attached two plots of my voltage source as measured with the 3458As. The setup was left to warm up for about ~5h and an ACAL was performed. The temperature and humidity in the room is controlled and maintained. The voltage source was primarily connected to the A3458A (Agilent with op02, last cal ~1.5y) and than leads have been switched to the K3458A (Keysight, without opt02, 2 months old) and than back again to the A3458A to capture some more data points. This is why there are small dips in voltage. Data was captured manually and I've left out the warm up phase of the reference from the plots. Reference is LTZ1000CH based. I think overall it works better than I've expected and with respect to the short time stability I was more thinking of ~30min. I do not have long term data, but I'm planning to repeat the measurements at some point.

[serg-el]

Multimeters need ~ 3 days to stabilize after a long shutdown.
The hard drops down for #2 are performing an internal auto-calibration.

[aronake]

What is the difference between 6581D and 6581T?

[Mickle T.]

R6581T - is a technological multimeter. It used only for ADVANTEST T2000 IC/SoC/ASIC test system, but not in metrology laboratories. R6581T don't have any official specs or manuals.

[aronake]

R6581T - is a technological multimeter. It used only for ADVANTEST T2000 IC/SoC/ASIC test system, but not in metrology laboratories. R6581T don't have any official specs or manuals.

Thanks! Except for different sales purposed, have anyone figured out what, if any differences there may be? Functions seems same. Could one have higher grade and tighter specification components or something?

[leighcorrigall]

R6581T - is a technological multimeter. It used only for ADVANTEST T2000 IC/SoC/ASIC test system, but not in metrology laboratories. R6581T don't have any official specs or manuals.

Thanks! Except for different sales purposed, have anyone figured out what, if any differences there may be? Functions seems same. Could one have higher grade and tighter specification components or something?

Hello,

Some of the questions that you are asking have already been discussed. It may be useful to search this topic for the answers before posting.

Regards.

[bob.cmhk]

While measuring one of my LTZ references @25°C in a thermal chamber over a longer periode with my R6581D I noticed that there is quite some t.c. visible. Therefor the internal temperature was logged together with the voltage readings. It turned out this t.c. is due to the meter itself and about 0.8ppm/K. A measurement of the t.c. of the internal reference itself revealed, that this t.c. is not due to the reference, but something else. Time to upgrade R6581!
Several improvements are documented here:
http://bbs.1ppm.cn/topic/23/advantest-6581磨机-zy的得意杰作
http://bbs.38hot.net/forum.php?mod=viewthread&tid=28698&highlight=6581

It just so happened that pipelie organized a group buy to replace R234 which boosts the zener voltage to 10V and R200 as part of the voltage to current converter.

R234 original: Alpha Electroncis SLD (7k34 X F 10k0)
R234 replacement: Vishay VHD200, hermetically sealed

R200 original: Alpha Electronics SS103F20k00 XF
R200 replacement: Vishay 300193ZT

The resistors arrived on friday so the weekend was used to replace them. Unfortunately there is no drop in replacement for R200, so the leads of the 300193ZT needs to be bend slightly into the correct position, while the VHD200 fit perfectly.
Now that the replacement is done it‘s time to let the meter settle with the new components installed, which is said to be about a month and see how this modification turns out.

Not to hide anything, removing the networks is a real pita and needs careful work and a lot of solder to heat up all pins equally. The vias are also very small and hard to clean. So unless you are well experienced in such work I would not recommend it, as you can damage a lot.
I use drill needles to finally remove last solder pieces inside the vias while heating the pad.
Also note that after that modification the meter needs calibration. To be continued ...

-branadic-

Thank Pipelie！Can we  organize a group buy again?

[leighcorrigall]

...

Thank Pipelie！Can we  organize a group buy again?

A group buy from Vishay got a lot more difficult. A MOQ of 25 is needed to place an order on their high end products. There might be a work around, but it's something new this year.

That said, I would be interested in a group buy myself.

[Kosmic]

...

Thank Pipelie！Can we  organize a group buy again?

A group buy from Vishay got a lot more difficult. A MOQ of 25 is needed to place an order on their high end products. There might be a work around, but it's something new this year.

That said, I would be interested in a group buy myself.

I would also be interested 

[little.tesla]

I'm interested as well 
Depending on price, probably two sets.

[aronake]

I have found a R6581D for sale with asking price of 700 USD. I could need some help from the expert panel here and people with experience using R6581 if this is good or not.

The seller know very little about multimeters, and not possible to test before buying except for telling him to make photos. With his limited insight is probably too much of an ask to tell him to measure voltage of a battery with it for instance. I asked for photos of the display as this typically is a critical component and asked for one photo each after pressing DCV, 2W and 4W resistance. I then got attached pictures and based on these I try to figure out the general health of the multimeter. Nothing was connected to the input of the meter.

DCV show 0.0000 in 1000V range. A bit odd that it do not auto-range down to a lower range and then show some voltage and could indicate a problem.

Question: Do R6581 remember manual range selection after a power down? if so that could explain.


2W resistance show Overload in 1000M ohm range. This seems to be what to expect, or do anyone typically see something else?

4W resistance is looks troublesome and broken.. 0.000000 ohm at 10 ohm range... Or is this some normal behavior?

If broken (which it kind of seems it is) do anyone recognize there problems?

The meter being broken may not have to mean that i do not want to buy it and it would be an interesting repair project. I have completed a couple already. Despite no official service manual people in this thread have made an excellent view analyzing it so that ought to help. At lease it have a very good display, so if repair fails that could be sold to someone who need a new.

Any other thoughts on buy vs no buy on thisone? Having read though this thread (and other) about R6581 I am well aware of its shortcomings.

[Mickle T.]

Question: Do R6581 remember manual range selection after a power down?

Yes it does.

2W resistance show Overload in 1000M ohm range. This seems to be what to expect, or do anyone typically see something else?

It's Ok.

4W resistance is looks troublesome and broken.. 0.000000 ohm at 10 ohm range... Or is this some normal behavior?

It's a normal behavior too.

[aronake]

Question: Do R6581 remember manual range selection after a power down?

Yes it does.

2W resistance show Overload in 1000M ohm range. This seems to be what to expect, or do anyone typically see something else?

It's Ok.

4W resistance is looks troublesome and broken.. 0.000000 ohm at 10 ohm range... Or is this some normal behavior?

It's a normal behavior too.

Awesome! Thanks a lot! And thanks for all work you have put into analyzing the 6581 and sharing your findings.

The meter still have original factory seal for warranty and was factory calibrated in 2006. So seems like it should be a decent purchase.

[thehater]

Would anyone have 2 VFDs for the advantest available?  We are working on 2 units, and they're broken

[TheDefpom]

Would anyone have 2 VFDs for the advantest available?  We are working on 2 units, and they're broken

I have been waiting for one to come up for about 10 months now, so far no luck, at this rate I will have to design a replacement display module.

[Mickle T.]

Advantest R6581 VFD can be easily emulated on a graphical LCD or OLED.