Advantest R6581 8.5 digit DMM mini teardown/repair

[Mickle T.]

In my case the rollover error is ~1 ppm with or without R8-R11, D003-D004, even without an input amplifier. So I think it's ADC INL  problem only.

[szszjdb]

Hi,  Mr.Kleinstein and Mr. Mickle.T ,

I still take that there has something happened around +8V to+4V ,that the bias become larger than others. That might also cause the bad INL arround 6V.  Attached FYI.

Further test about the gate of the MUX will be tonight.

Best Regards,
szszjdb

[Kleinstein]

I agree that there is still something wrong at positive voltages, higher than about 5 V. However this looks like it is about the same with AZ active and not.  The gate voltages did not help with that problem. For this increased leakage, there are quite a few possible paths:
Through D003/D004 - seems to be checked with voltages around U108
Guard driver not working well  -  checked with the gate signals
slow amplifier input - unlikely as this should not be voltage dependent
fast amplifier input - maybe if the fast amplifier does not work -> should be easy to check, reasonable working should be good enough
JFETs in the MUX  - some checks possible, e.g. for the Ohms_sense inputs, Hi_div.
   Other are a little tricky, but 1 nA * 10 K is still 10 µV and should be thus detectable on the low input.
   However leakage toward the gates hard to check. At least the -17 V level seems to be OK (not too negative).
  A possible check could be with changing the -17 V level for the gates a little (e.g. to -18 v or -12..-15 V): if this also shifts the leakage curve it would point towards gate leakage.

For comparison one could do a similar leakage test for the Ohms_hi input, to measure leakage at this input. With something like 1 MOhms or maybe 100 K with higher test current, one could have the cap from Ohm_hi to ground and than disrupt the connection to the resistor. However this would be only the negative side. Not sure if the DMM will record positive voltages on Ohm_hi and show the corresponding negative resistance values.

For the excess leakage below -5 V in AZ mode, this could be something like to much ringing on the DC / guard amplifier, or just a wrong / drifted value for C26 R083 (gate slow down). A smaller value for R083 might be worth a test. Also having the protection mode active could make a difference.

[TiN]

Would be direct current measurement helpful, using the electrometer for example?

[zhtoor]

maybe the 1nf cap used for these measurements acts funny at ca. 5-6 volts (maybe some kind of a diode action in junctions?)

best regards.

-zia

[Kleinstein]

One could do a direct current measurement with suitable pA/nA meter. However I don't think it would be very much different from the test with capacitor discharge. I don't expect an capacitor to behave that bad (except for an electrolytic maybe). The scale could be a bit off (maybe 10-20%) because of extra internal capacitance and at there might be a few 100 fA of current going to DA. External coupling could also add a bit of error in the initial phase when the hand is moved.

It might help to have some discharge curves from a well working meter. Even a well working meter might show enough input current to effect measurements with a higher impedance source. One does not know before checking it.

[eurofox]

I just did new testing by improving my LTZ1000 board, housing, twisting the cables to te meter, make an internal calibration and check again the polarity difference 

I get now the same value 

Not a bad meter and very happy with it 

Video on TiN server that show that I'm not telling "fake news"

[szszjdb]

Hi,  Mr.Kleinstein,
Thanks a lot!

Here are some photo of the gate test. The 1st is the detail of the ADIN overshot in +10v input. The 2nd to the 4th are the detail of PIN17/20 waveform at +10v input. It could be found some delay at the back of the AZ timing. The 5th to 10th are for the -10v input waveform of the ADIN/PIN14/PIN17/PIN20.  It also could be found some delay at the front of the AZ timing.  Do the delay imply some issues in it ?

I was testing a analog pcb from my friend ,which might have better performance than mine.

Best Regards,
szszjdb

[Kleinstein]

At first glance the overshoot in the first picture looks quite bad. However the possibly bad overshoot happens already well in the zero phase. So even if there is some gate current due to the gurad amplifier to slow, this would be only current at the ground pin. No it should not be real problem.
The more tricky one it the other slope. 
The 3rd picture kind of gives that information. The rising red curve in the later stage should reflect the rising guard amplifier signal and this looks a little suspicious. Possibly the slow DC amplifier might have reached it's slew rate limit and during this phase might take some input current.
A simlilar limited slew rate when going up is seen in the 9 th picture for pin 17. Here it is obviously following the guard amplifier going up at a constant rate (no extra RC with R17, making this more clear).

I would check the DC voltages in the slow amplifier input stage and maybe the effect of the trimmer. The input stage working with too much or too low current could be a problem. A slew rate of around 0.25 V/µs could also be still from a rather slow sample of U107. The first picture seems to show a slightly higher slew rate for the ADIN (= DC amplifier out).

A possibly interesting point could be the rising edge of ADIN (and maybe guard / pin 20) with an input voltage of +10 V and maybe around +4 V (so in the range where AZ still works OK).

[TiN]

Little bird show me photo of (later?) R6581T, which have DCV 300V PK MAX Input labels.
Seems like Advantest were not able to make up mind about 6581 specs for a while. 

Video on TiN server that show that I'm not telling "fake news"

Well, I watched it, and Mr. eurofox misled us, as meter he has is 6581T , not the real 6581.  No wonder Mickle.T. did not see any difference  .

[eurofox]

Little bird show me photo of (later?) R6581T, which have DCV 300V PK MAX Input labels.
Seems like Advantest were not able to make up mind about 6581 specs for a while. 

Video on TiN server that show that I'm not telling "fake news"

Well, I watched it, and Mr. eurofox misled us, as meter he has is 6581T , not the real 6581.  No wonder Mickle.T. did not see any difference  .

Maybe I'm wrong but R6581 or R6581T is the same basic unit, R6581T is not provided with the AC input and for some reason they limited the maximum DC input voltage.

In the menus all AC options are provided in the R6581T version.

I don't need AC input on a 8 1/2 digit meter

[TiN]

Idea is that they are the same, but in reality lot of issues with T-units we all see here leave us wondering if real 6581 actually meet the specified specs and 6581T is just reject units, or actually there are design changes/different parts that give 6581 higher performance.

[zhtoor]

Idea is that they are the same, but in reality lot of issues with T-units we all see here leave us wondering if real 6581 actually meet the specified specs and 6581T is just reject units, or actually there are design changes/different parts that give 6581 higher performance.

are the issues faced by Mickle T., szszjdb and you shared by eurofox, specifically the turn-over error?

regards.

-zia

[eurofox]

Idea is that they are the same, but in reality lot of issues with T-units we all see here leave us wondering if real 6581 actually meet the specified specs and 6581T is just reject units, or actually there are design changes/different parts that give 6581 higher performance.

are the issues faced by Mickle T., szszjdb and you shared by eurofox, specifically the turn-over error?

regards.

-zia

Advantest is a company with a very good reputation and I don't think that they classify high end instrument on the end of the assembly line, having myself managed a factory where electronic is assembled look ridiculous for me.

They probably sold more more R6581T to put in automatic testing configuration instead of the R6581 dedicated to lab. Reason why many of the R6581T have more service hours and prone to have defective parts, look to the display of most of the R6581T, some are on the edge of readability and this is a results of many service hours.

Sorry guy's my unit got a very bright display, perform wel on the polarity check. After a full external calibration I will be able to make accurate test of my unit.


eurofox

[TiN]

I do respect Advantest for doing the meter in first place, but in my book high-end instrument is not just a piece of hardware, but also documentation (anyone see any document about 6581T?), service procedures and reliability of the internal and external calibration procedures. Yes, long service hours affect instrument aging, stability and bring wear and tear issues. Actually that what is making high-end instrument so, ability to sustain it's tight specifications over the time and use abuse. One can calibrate "cheap" 6.5d DMM and 8.5d DMM to same initial accuracy, but over 1 year, you will see why 8.5d is priced 10 times more.

Single 7V/-7V point does not tell us much, sorry. You need linear source or a reference meter with proven linearity (hello, 3458A/5440/720A-732 etc), to verify issues that multiple people see.

Don't get me wrong, R6581T is nice meter to have, but saying it's on par with other industry proven 8.5d ones is bit over-excitement (unless we find out exact root cause of performance issues).

Or maybe all of the R6581T that multiple people have here are simply broken in various ways, yet pass external and internal calibrations and self-tests, giving a false pass and out-of-spec results.
Running external calibration on Ohms, with 10000.0011 Ohm standard and getting 9999.8791 ohm as result is bit unsettling .

Given simple statistics we should have at least one person here with R6581T which actually does meet R6581 specs? Maybe it's you? 

[zhtoor]

maybe the "T" in R6581T stands for Tolerance? 
btw. does anybody have a source for the display used in the meter?

-zia

[Mickle T.]

btw. does anybody have a source for the display used in the meter?

I found only this way: http://bbs.38hot.net/forum.php?mod=viewthread&tid=2014&extra=&page=1

[zhtoor]

btw. does anybody have a source for the display used in the meter?

I found only this way: http://bbs.38hot.net/forum.php?mod=viewthread&tid=2014&extra=&page=1

wow.

imagine the 10-year cal-cycle with a display like that !

-zia

[eurofox]

I do respect Advantest for doing the meter in first place, but in my book high-end instrument is not just a piece of hardware, but also documentation (anyone see any document about 6581T?), service procedures and reliability of the internal and external calibration procedures. Yes, long service hours affect instrument aging, stability and bring wear and tear issues. Actually that what is making high-end instrument so, ability to sustain it's tight specifications over the time and use abuse. One can calibrate "cheap" 6.5d DMM and 8.5d DMM to same initial accuracy, but over 1 year, you will see why 8.5d is priced 10 times more.

Single 7V/-7V point does not tell us much, sorry. You need linear source or a reference meter with proven linearity (hello, 3458A/5440/720A-732 etc), to verify issues that multiple people see.

Don't get me wrong, R6581T is nice meter to have, but saying it's on par with other industry proven 8.5d ones is bit over-excitement (unless we find out exact root cause of performance issues).

Or maybe all of the R6581T that multiple people have here are simply broken in various ways, yet pass external and internal calibrations and self-tests, giving a false pass and out-of-spec results.
Running external calibration on Ohms, with 10000.0011 Ohm standard and getting 9999.8791 ohm as result is bit unsettling .

Given simple statistics we should have at least one person here with R6581T which actually does meet R6581 specs? Maybe it's you? 

I don't have the "tools" to calibrate and test on the long run my instrument and I don't pretend mine is better but on the polarity check it perform well even better than some of your 3458A.

With respect to internal testing I have the impression the this test is very basic, it is to fast to be an advanced test and can probably not detect all issues of the instrument.

I google to try to find <> of the R6581 and R6581T but information is very limited.

Fortunately the front panel key system is "so logic" that you don't need a manual to use the instrument much user friendly than the 3458A.

[TiN]

I'm afraid this is going to some sort of holy war, which is not necessary, so I'll shut up after this post. Most of information about 6581/6581T differences discovered by Mickle.T who did great job unveiling the design and functionality of these. I was talking about any official documentation from Advantest about T-version. Without it we are left speculating and applying 6581 specifications to the 6581T model (as from what we seen, both meters share same platform).
If I calibrate 3458A versus external standard I get meter to meet it's 15 minutes transfer spec, no worries there. However my (broken) 6581T unable to meet 1 year spec right after cal? Hmmm  .

Fortunately the front panel key system is "so logic" that you don't need a manual to use the instrument much user friendly than the 3458A.

Just a matter of habit. I found 3458A very easy to use after programming macro onto numeric keys, so I can get 99% of modes and functions I want in click of two buttons.
Bling should not be the reason for performance tradeoffs, we are in metrology section after all

Looking forward to your ext.cal results.

[zhtoor]

I'm afraid this is going to some sort of holy war, which is not necessary, so I'll shut up after this post.

hey lighten up guys. no WAR only PEACE. (imagine that coming from a Pakistani  )

on a more serious note:

it looks like R6581T is the model specifically designed/selected to go into automated Test racks,
and i think all of the units are coming from some kind of testing setup.
has anybody evaluated the linearity using a fluke 720a kvd? maybe make some kind of a correction table in software?

best regards.

-zia

[eurofox]

I'm afraid this is going to some sort of holy war, which is not necessary, so I'll shut up after this post.

hey lighten up guys. no WAR only PEACE. (imagine that coming from a Pakistani  )

on a more serious note:

it looks like R6581T is the model specifically designed/selected to go into automated Test racks,
and i think all of the units are coming from some kind of testing setup.
has anybody evaluated the linearity using a fluke 720a kvd? maybe make some kind of a correction table in software?

best regards.

-zia

Nobody here have to intention for an "holly war" and I'm not religious at all 

I got in the past a HP3458A as well several years ago, I lost it because of Russian mafia, long story but true, according to police/court my money when to Russian account, don't ask me how is a long story and I prefer to forget it since this money is lost anyway

It seems that the few owner of R6581/T got problems with the instrument, it seems that there NO evidence based on information available that there is a difference between the "base" of the R6581 and R6581T, is the same firmware, on the R6581T all feature to manage AC input is available.

I was maybe lucky, the instrument that I scored is like new, no scratches, display is bright (visible on the video), battery was dead.
This instrument was never put in an industrial environment (coming from the industry I know how people behave with instruments) and I think it spend most of the time on a shelf unused.

[zhtoor]

It seems that the few owner of R6581/T got problems with the instrument, it seems that there NO evidence based on information available that there is a difference between the "base" of the R6581 and R6581T, is the same firmware, on the R6581T all feature to manage AC input is available.

well,

an obvious difference is the "T", i have seen japanese manuals and spec. sheets (again japanese) they refer ONLY to R6581 and R6581D, no T anywhere to be found.
a lot of component specification testing goes into manufacturing these 8.5d buggers, and a lot of variance in component selection is there, even the ref module
on an hp 3458a (seemingly simple LTZ1000 ref module) has a couple of versions available at *wildly* variable costs.

it does look like, that the specification for R6581T *is* different and it is a fallacy in my humble opinion to hold it against the R6581 or R6581D specs.

best regards, and always a pleasure to read *fantastic* pieces of knowledge shared by you guys.

-zia

[eurofox]

I did the same test of yesterday and get exactly the same values ....

Now yes we could say that my LTZ1000 board was changing and exactly the same way the R6581T ...

[eurofox]

It seems that the few owner of R6581/T got problems with the instrument, it seems that there NO evidence based on information available that there is a difference between the "base" of the R6581 and R6581T, is the same firmware, on the R6581T all feature to manage AC input is available.

well,

an obvious difference is the "T", i have seen japanese manuals and spec. sheets (again japanese) they refer ONLY to R6581 and R6581D, no T anywhere to be found.
a lot of component specification testing goes into manufacturing these 8.5d buggers, and a lot of variance in component selection is there, even the ref module
on an hp 3458a (seemingly simple LTZ1000 ref module) has a couple of versions available at *wildly* variable costs.

it does look like, that the specification for R6581T *is* different and it is a fallacy in my humble opinion to hold it against the R6581 or R6581D specs.

best regards, and always a pleasure to read *fantastic* pieces of knowledge shared by you guys.

-zia

You should show really facts not your impression based on a manual in Japanese because there is no specification on the T version, I have it as well and use software translators since I don't speak Japanese language.

When you switch it on it is "R6581D" why?

I wonder if you have a R6581 with or without T?

eurofox