T.C. measurements on precision resistors

[Andreas]

Hello,

as announced: UP805#1 as DUT with UP805#2 as reference resistor
I have also increased setpoint temperature of the reference resistor from 27.5 to 30 deg C
So temperature readings are more stable (within 0.01 K instead of 0.05 K) at the reference resistor.

UP805_12K5#1 delivery date 1510
T.C. +/-3 ppm max. from datasheet (typical +/-1 ppm/K)

21.06.2015: first measurement AC
22.06.2015: 2nd measurement AC
23.06.2015: 3rd measurement AC

hysteresis about +/-1.5ppm

LMS interpolation of 23.06.2015

A 0 =  4.03980761381591E+0000
A 1 =  7.41270259831732E-0001
A 2 = -2.85343759498466E-0002
A 3 =  3.63096053524801E-0005

So T.C. from LMS at 25 deg C is +0.74 ppm/K 

The "box" T.C. is around 0.71 ppm/K including noise
and around 0.68 ppm/K from LMS interpolation (without noise)

Again visible drift (3.99 ppm) during the 3 days

With best regards

Andreas

[Andreas]

Hello,

this time reversed: UP805#2 as DUT with UP805#1 as reference resistor (with 30 deg C setpoint).

UP805_12K5#2 delivery date 1510
T.C. +/-3 ppm max. from datasheet (typical +/-1 ppm/K)

24.06.2015: first measurement AC
25.06.2015: 2nd measurement AC
26.06.2015: 3rd measurement AC

hysteresis about +/-2 ppm

LMS interpolation of 26.06.2015

A 0 =  4.93434506507624E+0000
A 1 =  9.53465670990234E-0001
A 2 = -2.70096487231644E-0002
A 3 =  2.19612533635238E-0004

So T.C. from LMS at 25 deg C is +0.95 ppm/K  (slightly higher than #1)

The "box" T.C. is around 0.94 ppm/K including noise
and around 0.92 ppm/K from LMS interpolation (without noise)

Again visible drift (5.56 ppm) during the 3 days

With best regards

Andreas

[Andreas]

Absolutely not, this is the same problem for strain gauges for example. You can lower the excitation voltage using a little series resistor, see my schematic (This resistor = 1k).

Hello

sorry but in your schematic the 1K resistor also lowers the reference voltage.
So this resistor only reduces self heating of the resistors but will not be helpful since you cannot measure ratios above 0.5 with the LTC2440.
(you would have to put it in series to the reference resistor -> or increase the value of the reference resistor.)

Strain gauges measure only around 20mV of the 5V range so you cannot compare that.
Of course cou could use 3 reference resistors to measure one DUT to create a real bridge.

With best regards

Andreas

[MisterDiodes]

Andreas,
I have never, ever measured drift like that on any UltraOhm PWW (measured using a real bridge) over a year, let alone a few days.  That doesn't mean it couldn't happen, but you may still have something a bit off in your setup. 

Respectfully: You seem to be adding a lot of "nonsense digits" in your math (which can cause errors as well)- and remember that the ADC noise on the LTC2404 part you are using is not always perfectly random white noise - as you will soon be learning.  Also- watch your Vref.  It seems the drift data noise has gone up with increasing resistance - highly suspicious.  That means something is probably wrong in measuring technique in your measuring jig itself. 

For example, I don't take out a 25 foot carpenter's tape measure to measure a steel block to 0.01 micron accurately.  Which is kind of what you're doing here.  You want to keep your math resolution only to the accuracy of your measurement methods.  It is very, very hard to build a voltage-mode measuring jig that is stable & accurate to < 10ppm, let alone 1ppm.

Another item: you are not sure how well your reference resistor is temperature stabilized.  You think it is but it is very easy to be fooled by this one.  In fact we aren't absolutely 100% sure your test jig can even measure to some ppm accuracy to begin with.

Suggestion: You probably want to start using a resistance bridge for these measurements - looking for a "null" on a balanced bridge will give you much more stable results over time, and the measuring technique is much more suited to some x ppm measurements.  I know bridges are expensive but I know here in the states they are fairly inexpensive to rent for a few month's time.  Even a 3458a is really not the best tool for resistance measurements.  The best tool is a quality resistance bridge for what you are doing.

This doesn't cost a lot: I would suggest you pause and really look at your measurement technique accuracy and real resolution (accuracy and resolution are NOT the same thing) - maybe measure a 10k/ 10k LTC5400 ratio part for a reality check (these will get you a known ratio in ppm area guaranteed). LTC5400's will have more shot noise than PWW but are a cheap way to do a basic reality check on your system.

[MisterDiodes]

Andreas: You might have considered this already but -

Another "Gotcha!" to watch out for when using LTC24xx parts - remember your leakage current on the ADC inputs if you're not using a buffer.  That's going to sneak up on you if you start testing 12.5k resistors instead on 1k resisters.

See LTC2404 datasheet page 12 - even a 5k series impedance with the ADC input circuits  is going to start throwing off your measurements at ppm values with "Very strong temperature dependency".

Not sure if you have a buffer amp between your measuring jig test setup and the ADC, but this is yet another reason to be using a quality resistance bridge to investigate the stability of a resistor.

Just something to check.

[Andreas]

Andreas: You might have considered this already but - Another "Gotcha!"
Just something to check.

Hello MisterDiodes:

You are right I started with the 1K resistors without buffer.
(with 500 ohms resulting input impedance this is not a issue with the LTC2400)
But since I had some unclear offset issues due to rectification
of the input noise of the LTC2400 at the protection diodes
of the MAX4052A multiplexer I decided to use a ADA4538-1 buffer
between multiplexer and LTC2400. Of course the ADA4538-1 buffer
has a output filter 825R + some pF (carefully adjusted for linearity)
to prevent rectification of the LTC2400 noise on the output of the ADA4538.

So since 03.02.2015 I am using the ADA buffer.

I have never, ever measured drift like that on any UltraOhm PWW (measured using a real bridge) over a year,

So we have to regard the differences in measurement conditions:
You are using a clamp to contact the resistors.
I am soldering. I try to keep away the soldering heat from the body of the resistor
with 2 crocodile clips, but I cannot guarantee that there is no influence at all.

see also:
https://www.eevblog.com/forum/projects/vishay-bulk-foil-drift-after-soldering/msg445297/#msg445297

I do not know wether you have a room with controlled humidity.
I have none. So humidity changes from around 40-50% in winter to 60-70% in summer.
With several % change possible over a few days.

Perhaps with "precision lab conditions" the devices would perform better.
But since I intend to solder the resistors in the final cirquit
(because most users of resistors will not use any clamps in their cirquits) I will not change this.

By the way: where are your measurent results on T.C. and drift?
Up to now I have seen none from your side.

ADC noise on the LTC2404 part you are using is not always perfectly random white noise - as you will soon be learning.  Also- watch your Vref.  It seems the drift data noise has gone up with increasing resistance - highly suspicious.  That means something is probably wrong in measuring technique in your measuring jig itself. 

By the way I am using the LTC2400 (not the LTC2404 which relative similar) together with a MAX4052A/MAX4051A multiplexer.
Noise picture I have added below.
VREF is irrelevant in ratio measurements.
Ratio stability at half input voltage over a >8 hour period is shown to be around 0.1uV (giving 0.04 ppm error at 2500 mV).

Another item: you are not sure how well your reference resistor is temperature stabilized.

Stay on the carpet:
A few days before I wrote that I have improved temperature stability from below 0.05 deg C to below 0.01 deg C.
I monitor each resistor with 2 temperature sensors.
Please read more carefully.

maybe measure a 10k/ 10k LTC5400 ratio part for a reality check (these will get you a known ratio in ppm area guaranteed). LTC5400's will have more shot noise than PWW but are a cheap way to do a basic reality check on your system.

Don´t make yourself ridiculous the LTC5400 has worse specs in datasheet than a pair of Z201
or a typical pair of UP805 PWW resistors from the same batch.
And the HP3458A is still one of the best multimeter.

The only part that I trust on around ppm ratio stability is a LTC1043 capacitive divider.

Sanity check data attached:
Ratio check VREF/VIN at VREF divided by buffered LTC1043 2:1 divider.
1) Raw data of 8 hours = 170000 values (in mV)

2) Averaged values 1 minute as I use them for measuring to reduce noise. And 25 minute averages to show drift.

3) Classified raw data showing nearly perfect gaussian distribution.

4) Allan deviation showing stability over time in ratio mode. (better than 0.1 mV for large averaging periods)

compare that with the absolute voltage mode of a HP3458A:
https://www.eevblog.com/forum/projects/project-kx-diy-calibrator-reference-sourcemeter/msg592144/#msg592144
By the way there you can see: my ADCs (including compensated VREF) are even in absolute voltage mode
only around factor 5 less stable than a HP3458A if the 1 minute averaged values are used.
(please note that the measurements of 3458A are in "volt" whereas I measure in "milli volt").


With best regards

Andreas

[Edwin G. Pettis]

Andreas,

I'm working on an order so I'll keep it short at the moment.  Normal soldering will have no long term effect on my resistors but overheating can cause damage to the epoxy which should be quite visible, this could cause problems particularly if the lead assembly's anchor into the epoxy bobbin is damaged sufficiently.  Unlike the film/foil resistors which will can have permanent changes in resistance from soldering, mine do not under normal conditions, they will withstand over 150°C long term.  None-the-less, it is always good practice to keep high heat sources from precision components whether or not they can withstand such temperatures.

Humidity is another non-issue, under MIL-STD-202 humidity testing, there is no discernible change in resistance, the fact that film/foil resistors are made from very similar alloys is not the issue at all, it is the form of the resistor element that produces the sensitivity, the flat film/foil surface allows condensation on it and forms a parallel leakage resistance while the polyimide coating used on my wire prevents any condensation to form a parallel leakage path like the film/foil resistors.  In a PWW resistor, the leakage path would have to exist between the lead assemblies, many millimeters in length plus internal barriers, in the film/foil resistor the leakage path is in sub-millimeter lengths making humidity a much more significant problem (except for hermetics of course).  In PWW resistors that are not welded, humidity is an issue in the mechanical joint.

You are both correct and incorrect about the LTC5400 matched resistor chip.  True the absolute TCR is definitely worse than mine but their ratio performance (which is the main specification) is about the same as mine.  In a ratio setup, my resistors can outperform the LTC5400 but not by a lot, we are talking sub-PPM for both technologies here.  I also outperform them in lower noise significantly and on a price/performance comparison, a set of four of my resistors can be a little higher in cost but not a lot, The LTC5400 is a good bargain for certain requirements although the higher performing 'B' version is harder to come by and can be slightly higher in cost than mine, just depends on the required specs.  I believe MisterDiodes was suggesting the use of the LTC5400 as a ratio check since its ratio parameters are well known and guaranteed and the 'A' version is not difficult to obtain.

Since I am not familiar with all of the details of the various measurement setups that MisterDiodes is using, I will give him the floor for any details he wishes to give.

While you have posted photos of your test setups and schematics, it is difficult at best to 'trouble shoot' from a distance, there very well may be some very low level errors creeping into the measurements which are not consistent or vary with DUT value.  MisterDiodes is attempting the difficult task of accounting for possible errors given the measurements of the 12.5Ks (and others) you have posted, I can only say what characteristics my resistors have exhibited over years of production.  I am not saying it is impossible for a resistor to get through QC with some very small 'glitches', after all they are made by human hands but the manufacturing processes have been designed to eliminate resistors that are problematic before they are shipped and this has been exceedingly successful.

Again, there is no intent to take potshots at your work, we are merely suggesting that from the data you have posted, there is something apparently amiss, you are working in a difficult area of measurement.  What may have worked well at one resistance may have a problem at another resistance.  From time to time, I find errors have crept into my measurements and it can be difficult flushing out the culprit(s), it happens even in the best of cal labs.

Best regards,

Edwin

[MisterDiodes]

Andreas,

Edwin is correct, humidity effect should be minimal on PWW resistor.

I do not publish data on measurements with suspect results, and I can't publish data for my employers since I don't own the data that was collected at their expense.  The bottom line is that after hundreds of measurements on hundreds of different resistors, on a real resistance bridge (typically MI 6600 or ESI /SR '242 or similar) - if we saw a drift of over 5ppm in a few days on a PWW that would be considered a very, very suspect measurement, and probably something else wrong beside the resistor under test.  On a quality PWW like Edwin's, we wouldn't see that kind of drift in a year maybe.  Vishay's are a different story, but I'll save that for another day.

The 3458a is an OK instrument for voltage (if it is recent -adjusted- calibration), not the best for direct reading resistance UNLESS you are operating with a good reference standard.  Look at the specs.  Even then, a quality resistance bridge will have much, much better accuracy and repeatability than any 3458a, in any mode.  That's why you want to use a real bridge.

An LTC2400 is not a metrology grade instrument either.  Even with 17 million measurements, you haven't gotten a true average value yet - because the noise data is not white, nor is it random.  You'll find this out in time.

SUGGESTION:  Order an LT5400-1.  That will give you a couple pairs of 10k / 10k to work with with a guaranteed matching ratio of 1ppm TC (1ppm is an outlier, typical we see is maybe 0.2 to 0.3ppm TC).  Set it up in your jig.  Now take some measurements - you will see higher shot noise in diffused resistors but the matching is guaranteed.  On another day, take measurements again and this time adjust your Vref up or down a volt and watch what happens.

I know you think Vref doesn't matter, but you'll see why it does when you investigate closely.  You are comparing a S & H voltage on ADC inputs to the noise in Vref....and you're not even close yet.

Once you use a bridge, and see much more stable data, you'll see why 3458a's aren't used in calibration rooms to check resistance, even for relative values.  In fact at some places an old  '3456 is used for better accuracy (at less resolution ) than of 3458a...

You will certainly find 3458a's out on the semiconductor process lines for sure, but that's just used as a "daily driver" to get a solid 20 to 50 ppm accuracy that is dependable and fast on the testing platforms. For resistor checks in the ppm range, the instrument of choice and is almost universally used is a real, quality bridge.

[Andreas]

SUGGESTION:  Order an LT5400-1.  That will give you a couple pairs of 10k / 10k to work with with a guaranteed matching ratio of 1ppm TC (1ppm is an outlier, typical we see is maybe 0.2 to 0.3ppm TC). 

You are comparing a S & H voltage on ADC inputs to the noise in Vref....and you're not even close yet.

Hello,

I already have a couple of  LT5400-4 here. But I will not use them as a sanity check. Since they would have to be at least a factor 3 better than my instrument. I will only do a comparison to the DMSZ-Measurements of last year.

You seem really have no clue: A sigma delta converter does not have a S & H.
There is a continuously integration of input and VREF or GND.
The noise of the VREF (LTC6655 on ADC14) is a magnitude order lower than that of the LTC2400.

With best regards

Andreas

[bktemp]

SUGGESTION:  Order an LT5400-1.  That will give you a couple pairs of 10k / 10k to work with with a guaranteed matching ratio of 1ppm TC (1ppm is an outlier, typical we see is maybe 0.2 to 0.3ppm TC). 

You are comparing a S & H voltage on ADC inputs to the noise in Vref....and you're not even close yet.

You seem really have no clue: A sigma delta converter does not have a S & H.
There is a continuously integration of input and VREF or GND.

Andreas, you are wrong:
See LTC2400 datasheet:

Driving the Input and Reference
The analog input and reference of the typical delta-sigma
analog-to-digital converter are applied to a switched capacitor
network. This network consists of capacitors
switching between the analog input (VIN), ground (Pin 4)
and the reference (VREF). The result is small current spikes
seen at both VIN and VREF. A simplified input equivalent
circuit is shown in Figure 15.

[Andreas]

Hello,

UP805_12K5#3 as DUT with UP805_12K5#1 as reference resistor (with 30 deg C setpoint).

UP805_12K5#3 delivery date 1510
T.C. +/-3 ppm max. from datasheet (typical +/-1 ppm/K)

29.06.2015: first measurement AC
30.06.2015: 2nd measurement AC
01.07.2015: 3rd measurement AC

hysteresis about +/-1.5 .. 2 ppm

LMS interpolation of 01.07.2015

A 0 =  1.63629077103983E+0000
A 1 =  8.91505751881022E-0001
A 2 = -3.19358747995385E-0002
A 3 =  2.80114154455740E-0004

So T.C. from LMS at 25 deg C is +0.89 ppm/K (again within the typical spec of 1 ppm/K)

The "box" T.C. is around 0.83 ppm/K including noise
and around 0.80 ppm/K from LMS interpolation (without noise)

no significant drift (1.56 ppm) during the 3 days

With best regards

Andreas

[Andreas]

Andreas, you are wrong:
See LTC2400 datasheet:

Driving the Input and Reference
The analog input and reference of the typical delta-sigma
analog-to-digital converter are applied to a switched capacitor
network. This network consists of capacitors
switching between the analog input (VIN), ground (Pin 4)
and the reference (VREF). The result is small current spikes
seen at both VIN and VREF. A simplified input equivalent
circuit is shown in Figure 15.

And what has this to do with a S&H cirquit?

[bktemp]

Andreas, you are wrong:
See LTC2400 datasheet:

Driving the Input and Reference
The analog input and reference of the typical delta-sigma
analog-to-digital converter are applied to a switched capacitor
network. This network consists of capacitors
switching between the analog input (VIN), ground (Pin 4)
and the reference (VREF). The result is small current spikes
seen at both VIN and VREF. A simplified input equivalent
circuit is shown in Figure 15.

And what has this to do with a S&H cirquit?

switched capacitor = S&H

[Andreas]

This is no sample & hold.
This is a parasytic capacitor resulting from the input switches.

With best regards

Andreas

[Andreas]

Hello,

UP805_12K5#4 as DUT with UP805_12K5#1 as reference resistor (with 30 deg C setpoint).

UP805_12K5#4 delivery date 1510
T.C. +/-3 ppm max. from datasheet (typical +/-1 ppm/K)

02.07.2015: first measurement AC
03.07.2015: 2nd measurement AC
04.07.2015: 3rd measurement AC

hysteresis about +/-1.5 .. 2 ppm

LMS interpolation of 04.07.2015

A 0 = -1.52641348208928E+0000
A 1 =  9.16919772523320E-0001
A 2 = -2.54624953729614E-0002
A 3 =  5.40008996144842E-0005

So T.C. from LMS at 25 deg C is +0.92 ppm/K (again within the typical spec of 1 ppm/K)

The "box" T.C. is around 0.77 ppm/K including noise
and around 0.74 ppm/K from LMS interpolation (without noise)

drift -1.76 ppm during the 3 days

With best regards

Andreas

[Andreas]

Hello,

UP805_12K5#5 as DUT with UP805_12K5#1 as reference resistor (with 30 deg C setpoint).

UP805_12K5#5 delivery date 1510
T.C. +/-3 ppm max. from datasheet (typical +/-1 ppm/K)

07.07.2015: first measurement AC
08.07.2015: 2nd measurement AC
09.07.2015: 3rd measurement AC

hysteresis about +/-1.5 .. 2 ppm

LMS interpolation of 09.07.2015

A 0 =  3.13651864602711E+0000
A 1 =  9.47633786432885E-0001
A 2 = -2.89619803639599E-0002
A 3 =  3.58284202483480E-0004

So T.C. from LMS at 25 deg C is +0.95 ppm/K (again within the typical spec of 1 ppm/K)

The "box" T.C. is around 0.89 ppm/K including noise
and around 0.87 ppm/K from LMS interpolation (without noise)

drift 3.76 ppm during the 3 days

With best regards

Andreas

[Edwin G. Pettis]

Andreas, #570

Referring to Fig. 15 in the LTC2400 data sheet:

The Ceq referred to in the figure is not a parasitic capacitor, it is the equivalent of the distributed capacitors in the "internal switched capacitor network".

Quoting from the data sheet:  The analog input and reference of the typical delta-sigma
analog-to-digital converter are applied to a switched capacitor
network. This network consists of capacitors
switching between the analog input (VIN), ground (Pin 4)
and the reference (VREF). The result is small current spikes
seen at both VIN and VREF. A simplified input equivalent
circuit is shown in Figure 15.

The 'schematic' of Fig. 15 is an equivalent, not actual schematic of the switched capacitor network.  Internal capacitors on dies are very small by physical requirements, even a few pF takes up a lot of room on a die, therefore they must be small.  In this case, the sampling capacitor, identified as Ceq, is only 10pF where as on the LTC1043, the sampling capacitors are external (allowing for much lower clocking frequencies) and much larger values, i.e 1.0uF, obviously a 1.0uF capacitor cannot be integrated onto an IC die.

Again, quoting from the data sheet:  The key to understanding the effects of this dynamic input
current is based on a simple first order RC time constant
model. Using the internal oscillator, the LTC2400’s internal
switched capacitor network is clocked at 153,600Hz
corresponding to a 6.5µs sampling period. Fourteen time
constants are required each time a capacitor is switched in
order to achieve 1ppm settling accuracy.  Therefore, the equivalent time constant at VIN and VREF
should be less than 6.5µs/14 = 460ns in order to achieve 1ppm accuracy.

Please note the use of the term 'sampling' in the above paragraph, by definition, a switched capacitor network is a sample and hold circuit without doubt.  Much older integrating ADCs also used switched circuits but the signal was integrated for a complete measurement cycle, not truly sampled even though the input was not being integrated constantly because of the auto-zero cycle, it did not constitute a sample/hold circuit.

[MisterDiodes]

Andreas,
As others and I have pointed out, please see LTC2400 data sheet pg 22. You have a sample and hold as 5k switch feeding 10pF cap, as well as Vref.  That 10pF is not a "Parasitic" on inputs, that is a purpose-built capacitor on the die and functions a "hold" capacitor during conversion.  It is temperature sensitive as well as the input switch network (Noted in Linear Documentation).

You are assuming VRef doesn't matter, but it does very much matter:  You are comparing a voltage (+ noise) out of your resistor divider at time t1, to a voltage Vref (+ noise) at time t2.  Two different voltages measured at two different times. At ppm levels, this is comparing apples to oranges.   That's why your data will have non-random artifacts generated by switching freq, a 5k and 10pF RC, and the difference of Vin vs VRef.  Those all play together to generate pesky math averaging errors.  Yes the data looks like it has a Gaussian noise curve, but its not truly random.  Watch what happens when you start your Vref at 3V and raise it to 3.5V.  Then start at 4V and lower it to 3.5V.  You will soon learn what happens in your averaging.   

If VRef weren't important, you should be able to use a slow sine wave for VRef.  That's another interesting experiment for another day.

Also notice the INL and data noise values changes with your VRef.

That 5k resistance in 2400 and 10pF hold cap are very temperature sensitive as you'll notice in datasheet.  And humidity sensitive as well, except that isn't explained nearly as well.  Test that yourself and you'll see.  The LTC2400 was designed long ago primarily for weigh scales / temp sensors etc and is very good to say 20~15ppm.  Using it to divine voltages in 1ppm area is theoretically possible, but very hard to achieve in real life as you are finding out.

If you're not using proper guarding, you might want to read up on that also.

SUGGESTION:  Please try a LT5400-1 in ratio mode to verify your test setup.  And then try a 3V Vref vs 5V Vref and watch what happens.  Any LT5400-1 A or B grade will get you in the world of <1ppm TC ratios right off the shelf (guaranteed), and then you can work on your test setup to see what is going wrong.

While it is possible that your PWW resistors are drifting huge amounts, Long Experience tells me you probably have something else wrong.  We use UltraOhm resistors here, and 13k values, and have never seen that kind of drift in a few days.  Maybe over the course of a year maybe 3~5ppm maybe.  Not that much in a few days...never-have we seen that.

In the end, if your intention is to use these resistors with LTZ1000, it hardly will make any difference anyway if the heater ratio resistors have similar TC and in the same direction.  The LTZ does a very good job at attenuating any resistor drift - down to the point you can't measure it with any equipment owned by average hobbyist or even industrial equipment - at least not within the range of drift inherent in the measuring instrument itself.

Again:  In the ppm world of resistance measurement, the method universally used is a resistance bridge for best accuracy - even when comparing relative resistor values.

[Andreas]

The Ceq referred to in the figure is not a parasitic capacitor, it is the equivalent of the distributed capacitors in the "internal switched capacitor network".

Please note the use of the term 'sampling' in the above paragraph, by definition, a switched capacitor network is a sample and hold circuit without doubt.  Much older integrating ADCs also used switched circuits but the signal was integrated for a complete measurement cycle, not truly sampled even though the input was not being integrated constantly because of the auto-zero cycle, it did not constitute a sample/hold circuit.

Hello,

ok so I obviously missed something from the data sheet.
Sorry for inconvenience.

But what is the consequence:
I have either to use low ohmic resistors (< 1K)
or I have to use a precision buffer which:
- has low output impedance
- is not disturbed by the "switching noise" of the LTC input.

And that´s exactly what I do with my ADA4538 buffer (with filtered output).

With best regards

Andreas

[Edwin G. Pettis]

No inconvenience at all, it is just another name for the same function done in a different manner than a 'classic' sample and hold circuit.

[Edwin G. Pettis]

Andreas,

I can't find anything on a ADA4538 except that it is an alternator and I don't think that is what you are using.

[Andreas]

Andreas,

I can't find anything on a ADA4538 except that it is an alternator and I don't think that is what you are using.

Hello,

of course it is the ADA4638-1
http://www.analog.com/en/products/amplifiers/operational-amplifiers/zero-drift-amplifiers/ada4638-1.html

with best regards

Andreas

[ltz2000]

Maybe the internal layout is similar to the Fluke 742A.

There are high resolution internal photos of the Wekomm resistor in this Daves video, from 5:17.



I don't know how to extract the photos from the video...

[plesa]

Maybe the internal layout is similar to the Fluke 742A.

There are high resolution internal photos of the Wekomm resistor in this Daves video, from 5:17.

I don't know how to extract the photos from the video...

The enclosure is Rolec ALuPlus 100 http://www.rolec.de/en/aluPLUS/196.100.000/AP100.pdf
And resistors are VHA518-7   http://www.vishaypg.com/docs/63625/63625.pdf
VHA518 datasheet http://www.vishaypg.com/docs/63120/hzseries.pdf
According to information from Vishay datasheet VHA518-11 with PMO seems to be better option. Maybe Wekomm needs to to make some selection from Vishay batches...
What is interesting, that on first photo are wires and resistor soldered and on third photo wires are crimped to resistor.
Wire seems to be PTFE insulated and stranded silver coated. I have no idea why they used stranded wire instead solid copper. Any ideas?

[Dr. Frank]

And resistors are VHA518-7   http://www.vishaypg.com/docs/63625/63625.pdf
VHA518 datasheet http://www.vishaypg.com/docs/63120/hzseries.pdf
According to information from Vishay datasheet VHA518-11 with PMO seems to be better option. Maybe Wekomm needs to to make some selection from Vishay batches...
What is interesting, that on first photo are wires and resistor soldered and on third photo wires are crimped to resistor.
Wire seems to be PTFE insulated and stranded silver coated. I have no idea why they used stranded wire instead solid copper. Any ideas?

According to the designer of these standards, the resistive element is a special construction, no VHA518, although it might be in the case of this type.
PTFE insulation is very important for low leakage, as are the special and very expensive banana jacks.

I found my notes of that talk, so silver coated wire should be better than pure copper because of thermal voltages.
Anyhow, as you usually use Offset Compensation, this effect can be mitigated in any case.

Frank