Ultra Precision Reference LTZ1000

[Dr. Frank]

The goal of the design of the reference system was to calibrate a HP34401A on all DCV ranges, and a Fluke 332B, including its linearity to 1ppm of decks A, B; the uncertainty originating from one LTZ1000 reference only .

Here's the setup, as intended, for the case that the 34401A is the DUT.

The 332B simply serves as a high stability volt source (uncalibrated).

Its output of roughly 100V were first divided precisely (<0.5ppm) by 10.000  by a Hammond type divider (analogue to Fluke 752A).
The divided 10V were compared to the LTZ1000 box, which delivers different reference voltages, i.e.
7.1479788V for Ref_1, 7.1762318V for Ref_2, directly from both of the LTZ1000, 7.000000V divided very stable from one of the LTZs.

All three voltages can be very precisely (< 0.2ppm) amplified by 10/7, for the purpose to have usable calibration voltages for the instruments, in this case to 10.00000V.
(The 10,251760V and 10,211398V would also be accepted by the 34401A, and would have lower uncertainty).
 
The 1µV sensitive DVM compares the 10V reference to the divided one from the 332B, so that latter output is tuned to exactly  100.00000V.

Those 100.00000 V are fed into the HP34401A, for calibration.

All Cardinal Points: 1kV, 100V, +/-10V, 1V, and 100mV can be obtained precisely by different set ups of the 10:1 / 100:1 divider together with the LTZ box.


Later on, a HP3458A and a Fluke 5442A were available, so the stability and uncertainty could be checked.

The additional requirement then was, that the LTZ1000 references should be more long term stable than the 3458A. (1ppm/yr. vs. 8ppm/yr)

The 100:1 divider has a lower uncertainty on measuring 1000V (1ppm) than the 3458A (12ppm), because latter one has no compensation for the heat effect of its internal HV divider.

- to be continued -

Frank

[Andreas]

Here's the setup, as intended, for the case that the 34401A is the DUT.

The divided 10V were compared to the LTZ1000 box, which delivers different reference voltages, i.e.
7.1479788V for Ref_1, 7.1762318V for Ref_2, directly from both of the LTZ1000, 7.000000V divided very stable from one of the LTZs.

Hello Frank,

nice gear.
I hope you will describe further how you get (with which parts) the precise division ratios.
By the way: do you know the (traceble) uncertainity of your references against PTB standards?

Another gadet that is missing in the datasheet cirquit of LTZ1000 is the bandwith limitation of the current source amplifier like C2 + R12 in the jj3055 cirquit ver3.gif in the posting above. Without this R/C-Combination the reference voltage output will be unstable with capacitive loads above around 1nF. (The current source will have heavy oscillations).

Also a output buffer is missing. Although introducing additional errors, the advantage is that with a short cirquit to ground at the output the setpoint voltage of the heater is not affected. The heater will go to the maximum possible temperature on a short introducing hysteresis or even ageing effects to the chip.

With best regards

Andreas

[Dr. Frank]

I'd like to ask for some patience..  practical stuff will be delivered definitely.

But first, here’s the summary of the requirements of the reference system, already with some instructional comments.

1) The Reference System should be capable of calibrating a HP34401A on each DCV range, over years. The HP34401As 24h specification is 15ppm for 10V, 20ppm on 1000V, 30ppm on 100mV.

For Good Metrological Practice, an uncertainty of <2ppm is required for each range of the Reference System.

That means, if the uncertainty (relative to SI) of the LTZ1000 output is known (i.e. calibrated once), all transfers to the different ranges must be done below that limit.
Also, all drift parameters should contribute in total below that limit.

2) The stability over time of a single LTZ1000 reference should reach realistically <= 1ppm/year:
LT specifies for the naked chip typ. 2µV/sqrt(khr.) @65°C, that’s 0.8ppm/yr.

This stability requirement (aging or deterioration) is most important, as it is not reversible, and cannot be mitigated or hardly be compensated (at most by elaborate trend analysis), like other instabilities.

3) The stability over temperature of the LTZ circuitry should reach realistically < 0.2ppm/K.

This instability is reversible, by returning to the nominal temperature, and can be cancelled statistically or even mitigated.
For that purpose, the reference system has to be operated in a stabilized / controlled environment. Temperature changes must stay within a few 1/10°C during 10 min, or over the measuring period (hours), and for long term stability measurements, the room temperature must be reproducible to e.g. +/- 2°C.

Otherwise, drift measurements on sub-ppm level are not possible.

Btw.: All other sources of similar instabilities, e.g. thermo electrical/mechanical force induced, humidity/leakage current, pressure, gravity, and so on, have to be analysed in value, relative to the ageing drift, i.e. if it's worthwhile to cancel them.
Like the temperature coefficient, those drift sources lead to spurious / reversible modulations of the reference output "only".   

4) Simply for convenience, the raw 7,2V output of the LTZ1000 has to be attenuated and trimmed precisely to a plain value of 7,00000V.
This secondary output should be very stable over time and temperature.
This transfer might be done by a 34401A with 2ppm uncertainty (linearity error), or with a 720A (<0.2ppm), or a 3458A (<0.05ppm). 

5) All reference outputs shall be amplified and buffered with low impedance by an exact factor of 10/7 to around 10V, so that the 34401A will accept this for calibration.
This amplification factor may be auto-calibrated at any required time with an uncertainty of < 0.2ppm, so that the stability / uncertainty of the LTZ1000 is maintained in the 10V output.

6) A decade divider with precise 10% steps, 0.1ppm nonlinearity, should provide a means for calibrating the linearity error of a Fluke 332B to 1ppm.

7) The experimental verification of the stability of the LTZ reference and the complete system is required, due to Good Metrological Practice.
A theoretical model about the instabilities is not sufficient on its own.

Therefore, a measurement system is required which is capable to perform sub-ppm comparisons, of the DUT against a more stable standard, or against a group of equivalent stable references.

8.) A Reference Divider should provide precise ratios of 10:1 and 100:1 with uncertainties of < 0.2ppm and < 0.5ppm of output, latter one for a burden of 1kV also (like a Fluke 752A)

- To be continued – 

Frank

[Dr. Frank]

It’s teardown time; we’re on EEVBlog, aren't we?

First picture shows the double LTZ reference board.

It was intended as a Prototype design, to first check its characteristics, and some additional features.
It worked so well, that I left it in this state.

The PCB is single sided, and the LTZ1000 sits on the solder side (pic. 2).
That makes the thermal shielding of all solder junctions and the isolation of the LTZ much easier.
(Remember: Other ref. PCBs contain solder junctions on both sides and they are well exposed to air flow, especially in the 3458A, also in the Datron units)

The solder side is thoroughly cleaned with a strong solvent / PCB cleaner, and afterwards sealed with highly isolating plastic spray.
This method does not harm the other components.

The complete PCB is tightly assembled (with screws) on a polystyrene box, with additional small cavities for the LTZs, and the volume is filled with an additional foam cushion (pic. 3).
This easily fulfills the requirement that no air flow will affect the solder junctions, and also that the LTZ and its junctions are all on the same temperature, heated by its own power dissipation.

I have chosen an LTZ1000 (no A version!), so I could select 45°C as stabilization temperature for lowest drift.
That implies R1 = 1k, R2 = 12k. The rest of the LTZ circuitry is copied from the LT datasheet.
The LTZ1000A, as in ‘babysitters’ design, would require 10°C higher temperature, therefore 12.5k, and therefore 2 times higher drift rate, theoretically.

I have chosen wire wound resistors, because the necessary values were available instantaneously from stock, were reasonably cheap, and have low enough T.C. and specified long term stability of 25ppm/yr, no load. No load usually means, up to 10mW. Even the 120 Ohm resistor has 2.5mW only.

The OP07 simply were in my inventory, there’s no further idea behind using single opamps.

In the current budget schematic, (pic. 4) you see, that there are about 5mA flowing out of the negative side of the LTZ. That would have led to level shift on the negative line, especially as I connected both LTZs to one common GND.
So I added another OP07 to sink the negative current, that’s some sort of current cancellation, which is incomplete (and to be improved), but also used in the Fluke references (732B, 5720A).

That required a negative supply, so I have split the 15V to +12V/-3V, visible on the 2nd PCB.
The LTZ circuitry is effectively running on 12V only.   

The direct LTZ1000 output is divided by a resistive divider trim able precisely to 7,0000V.
This divider is set up by ordinary wire wound resistors also, i.e. T.C. ~ 3ppm/K, drift < 20ppm/yr., but the output is much more stable, i.e. 30 times more stable than those numbers.
That’s a metrological trick, I think, based on the fact, that at a division factor close to 1.0, all drifts will be suppressed strongly. The error calculus will be presented later.

Both LTZ1000 and one 7,000V outputs are fed directly to front jacks – gold plated Cu plugs.

The cables are Teflon isolated, to avoid leakage.

The precision switches allow choosing any of those 3 voltages, for amplification by a ChopAmp 7652 exactly by a factor of 10/7.

This is accomplished by 11 Vishay metal foil Z201 resistors, 49k99, which can be equally trimmed to 50k000 by a Wheatstone bridge, and from the front panel in Cal Mode.
The divider chain can be identified in pic. 1, on the right side, all those green components.
The linearity of this divider is 0.1ppm. This is acting like the first decade of a 720A Kelvin Varley divider. See 5th picture.

Therefore, the theoretical error of this 10/7 ratio transfer is 0.14ppm, which does not spoil the uncertainty/stability of the LTZ1000.

This is the 2nd metrological trick, as all other references (732A/B, 7001, etc) suffer on stability because of their fixed resistive amplification.

This divider delivers decade steps from 1 to 10, with an uncertainty of 0.1ppm of input , i.e. +/- 1µV on each output tap. (See 720A datasheet).
Therefore, 1V output is uncertain to 1ppm.

1V and buffered 10V are directly fed to output jacks, also all voltage steps can be selected by a 2nd switch.

Last, the calibrated 11 step divider can be used isolated for precision ratio measurements.

- To be continued -

Frank

[fmaimon]

In the current budget schematic, (pic. 4) you see, that there are about 5mA flowing out of the negative side of the LTZ. That would have led to level shift on the negative line, especially as I connected both LTZs to one common GND.
So I added another OP07 to sink the negative current, that’s some sort of current cancellation, which is incomplete (and to be improved), but also used in the Fluke references (732B, 5720A).

That required a negative supply, so I have split the 15V to +12V/-3V, visible on the 2nd PCB.
The LTZ circuitry is effectively running on 12V only.   

Can you explain this part of the circuit? From the schematic, I see that you are forcing the bottom of the 120 resistor to 0V, using the opamp, but isn't this same current still going to ground through the opamp?

Maybe I'm missing the point on what's ground in your circuit The ground symbol (upside down triangle) is the negative of your power supply (-3V of the split supply)?

And what is that diode marked with an * doing?

The direct LTZ1000 output is divided by a resistive divider trim able precisely to 7,0000V.
This divider is set up by ordinary wire wound resistors also, i.e. T.C. ~ 3ppm/K, drift < 20ppm/yr., but the output is much more stable, i.e. 30 times more stable than those numbers.
That’s a metrological trick, I think, based on the fact, that at a division factor close to 1.0, all drifts will be suppressed strongly. The error calculus will be presented later.

I'm looking foward to see this error calculus.

The precision switches allow choosing any of those 3 voltages, for amplification by a ChopAmp 7652 exactly by a factor of 10/7.

This is accomplished by 11 Vishay metal foil Z201 resistors, 49k99, which can be equally trimmed to 50k000 by a Wheatstone bridge, and from the front panel in Cal Mode.

How do you select the cal mode  50K resistors? How close does they have to be to each other and how (why?) do you need the balance trimpot? As I see it, as long as all resistors (R1..R10) are trimmed to be the same is enough.

Felipe

[Dr. Frank]

Can you explain this part of the circuit? From the schematic, I see that you are forcing the bottom of the 120 resistor to 0V, using the opamp, but isn't this same current still going to ground through the opamp?

Maybe I'm missing the point on what's ground in your circuit The ground symbol (upside down triangle) is the negative of your power supply (-3V of the split supply)?

And what is that diode marked with an * doing?

I'm looking foward to see this error calculus.

How do you select the cal mode  50K resistors? How close does they have to be to each other and how (why?) do you need the balance trimpot? As I see it, as long as all resistors (R1..R10) are trimmed to be the same is enough.

Felipe

Hi Felipe,

yep, the designators are not consistent, sorry for that.

I have +12V / GND / -3V; the triangle is -3V.

GND is the negative reference potential for each LTZ circuit.

My problem was, that I had to connect GND from the 1st and the 2nd LTZ circuit at the negative output jack, to have the opportunity to amplify each by the Chopper.

So, the negative sink currents of ~5mA for each ref. would have flown over those cables, which would have caused a (constant) voltage drop of several ppm, between the LTZ and the negative jack. So I reduced these currents and also the voltage drop by sinking them directly over the additional opamps, and only 300µA will flow to the negative jack.

Still, its not perfect.

Normally, you would sink the  currents directly at the LTZ negative output to the power supply, and would sense the negative potential with a separate tap, currentless.

Even then, you can only connect those negative sense lines without problems, if each LTZ circuit would have its own floating supply.

It was a first attempt, and I would redesign that , deleting the additional opamp, or would add several more precise current cancellation circuits, see 732B manual, or 5720A schematics.


The additional diode prevents reversing the voltage over UBE of Q1, by the additional OpAmp, which would otherwise cause the BE diode to go into Zener mode, which would destroy the chip.


The Wheatstone bridge is nothing special. It's simply a copy of the 720A bridge circuit, in its manual you may find a perhaps better description of the calibration  process.
In brief: The 50k resistors in the right leg are 0.1% wire wound, and both are trimmed by additional fixed resistors to nearly equal values (not drawn in the schematic).
The trim pot between them is the fine balance for exact match of upper and lower 50k.

The complete bridge has to be balanced to zero in each position, i.e. also if the positions of the 50k resistors in the right leg are reversed.
For that, you have to trim both pots, the trim of the DUT , and also the balance pot.
 
Only then, the resistor under test (R1..R10) exactly matches the reference resistor R11.
With 20V bridge excitation, and 0.5µV zero indication, the match is to 0.1ppm

Frank

[Andreas]

The OP07 simply were in my inventory, there’s no further idea behind using single opamps.

Frank

Hello Frank,

When comparing the datasheets of OP07 and LT1013 from the LTZ1000 datasheet the main difference is the large signal amplification.
LT1013 has a order of magnitude more amplification than the OP07.
Especially for the current regulating loop a high amplification will reduce the steady state current error.
I dont know how large the remaining error is. But since the open loop amplification is temperature dependant I would prefer using a OP with high open loop amplification like LT1013, OP177 or something else.

By the way. At the moment I am comparing noise levels of different ADCs against the datasheet value of the HP3458A.
If I understood it right the HP3458A has 10e-8 rms noise with a integration time of 2 seconds (100 NPLC).
So this would be +/-0.3uVpp or 0.6uVpp in the 10V range. Can you confirm this value? What is your experience from practical measurements?

With best regards

Andreas

[Mickle T.]

Since about 2008, Fluke changed the design of the reference board in the 8508A DMM. In early versions were installed an LT1413A and LTZ1000. Now it is AD823A, LT1150 and LTZ1000A.

[babysitter]

Dang Fluke, we all did everything wrong by selecting unsuitable chips...

[Andreas]

Since about 2008, Fluke changed the design of the reference board in the 8508A DMM. In early versions were installed an LT1413A and LTZ1000. Now it is AD823A, LT1150 and LTZ1000A.

Do you have a cirquit diagram or at least a hint where the Chips are used? (heater section, current regulator, output buffer?)

I am guessing that the LTC1150 will be used as current regulator for the zener since he has a very high open loop amplification and very low offset longterm drift. The AD823 will be probably used as current buffer since the LTC1150 cannot deliver the 5mA which are needed.

Nice picture: Is the board under the isolation slotted or without slots?

With best regards

Andreas

[Mickle T.]

I don't have any information about the new Fluke reference. All pictures taken from Web.

[Dr. Frank]

Hello Frank,

Especially for the current regulating loop a high amplification will reduce the steady state current error.
I dont know how large the remaining error is. But since the open loop amplification is temperature dependant I would prefer using a OP with high open loop amplification like LT1013, OP177 or something else.

By the way. At the moment I am comparing noise levels of different ADCs against the datasheet value of the HP3458A.
If I understood it right the HP3458A has 10e-8 rms noise with a integration time of 2 seconds (100 NPLC).
So this would be +/-0.3uVpp or 0.6uVpp in the 10V range. Can you confirm this value? What is your experience from practical measurements?

With best regards

Andreas

Andreas,

I had to take a time-out, but now, several chapters will follow..

Concerning the noise of the 3458A, compared to different stable DC sources, here's a measurement, up to 1s (NPLC50).
You can see, that those sources come near the theoretical limit of the 3458A, but factor 2 higher.
I simply have calculated the statistics of a set of measurements and identified the variance as RMS noise. Perhaps it's Vpp instead, so a factor of 1/ 2.8 would be possible.

Perhaps a measurement on a stable battery would give better results.. I'll see later today, the box is running currently.


As the temperature stabilization of the LTZ is very sensitive, a high open loop gain may lead to instabilities. I'll see, just ordered 5 EA of LTZ1000, LTC1013 and LTC1052 from LT..

Frank

PS: On the 10V range, I've just measured one Weston cell, and the LTZ refs, NPLC 100, 25 samples each:
 
Mean= 1,0186060V, sigma= 195nV (output is drifting)
Mean=7,1479820V, sigma=146nV
Mean=7,1762512V, sigma=120nV

Therefore, I would judge, that NPLC100 really gives +/-1 digit stability on 8 1/2 length.

[Mickle T.]

HP 3458A 10 V noise (taken from volt-nuts thread on the bbs.38hot.net)

[babysitter]

Do I spot a certain LTZ1000A trace in Dr. Franks diagram? 

My reference circuit will be shipped to another volt-nut this week (off-band for this forum) for further comparison. Also I will start building a LM399AH-based reference at work, or to be exact, prepare a design to be made by our trainee, will be used for performance tests of our 34401A at work (we have a scheme of buying fresh calibrated 34401A in place of a calibration sometimes and do a in-house comparison as performance/fitness check for the older ones, and some smaller fitness tests inbetween)

BR

Hendrik

[Dr. Frank]

The stability of the output voltage over time and temperature is determined mainly by the LTZ1000 chip, and the 5 resistors, R1-R5 (see datasheet).

Let’s first consider ageing (stability over time).

For the LTZ1000 chip and all of the resistor technologies there exists an exponential law over temperature, (Arrhenius), saying that ageing rate doubles every 10°C.

Such dependencies from operating temperature are often encountered in deterioration of material, e.g. the decay of LED luminosity and the life time of light bulbs, which is always driven by the operating temperature. The underlying process is the increase of defects in the crystalline structures, or the acceleration of diffusion processes. Therefore, reducing the operating temperature will reduce the ageing rates.

LT claims typical ageing of 2µV/sqrt(t/1000h) @ 65°C, that gives 0.8ppm/yr.
By lowering the temperature to 45°C, typical values of 0.2ppm/yr. may be achieved.
Therefore, the LTZ1000 version has to be used, with R4=12k, R4=1k. The reference then can be operated at an ambient temperature up to 35°C only.

Spreadbury (1) and also Pickering (2) have confirmed this experimentally for the LTZ1000 reference chip. The typical drift rates really apply in most cases. Intermittent operation of the LTZ1000 also will reduce its annual drift.
In the HP3458A the potential ultra-stability of the LTZ1000A is spoiled, by running it at 95°C, which leads to an 8 times higher ageing rate (continuous operation).

The 5 resistors also have to be operated on lowest possible temperature. This is fulfilled by Tamb. < 35°C and limitation of the self heating effect: Pmax. < 10mW. For those conditions, a Shelf Life ageing parameter is specified for precision resistors. In the LTZ1000 circuitry, R4=12k, 4mW, and R1 = 120, 2.5mW fulfill this requirement.

Additionally, there exists an influence from oxygen (and other reactive gases) and humidity on the crystalline structures, causing a sort of corrosion on resistors. Molded components, and to a lower degree conformally coated components suffer from that. Therefore all high quality / precision resistors as Thin Film, Wire Wound and Metal Foil types have similar specified ageing rates of 20...35 ppm/yr., (typical or maximum).

The LTZ1000 chip is hermetically sealed; therefore oxygen and humidity have no influence.

The resistors influence the reference voltage by changing the temperature set point (R4, R5), or by changing the current through the zener (R1).
The instabilities of R1, R4 and R5 are attenuated by a factor of 100; R2 and R3 by 300 and 500, respectively (see datasheet).
If all five resistors have the same instability values (over time or temperature), the total impact will be 0.035 times their instability.
Resistor ageing of 20ppm/yr. will add 0.7ppm/yr maximum.

Only the hermetically sealed, oil filled types (e.g. VHP202Z) give a big advantage. Their rate is typically 2ppm/6yrs., and therefore will add about 0.02ppm/yr only.
In picture 1 you’ll find long-term stability monitoring of 3 EA of my 5 VHP202Z. After 2 years, they really remain within < 0.5ppm of their initial value, so that is obviously no fake advertisement.
(Remark: The measurement stability was improved also during that time.)
 

Second, the temperature coefficient can be calculated in the same manner.

As the LTZ1000 chip is heated and thermally isolated, there is no further mechanism that would directly change its output voltage by change of ambient temperature.

The resistors will contribute with 0.035 times their own T.C.

Thin film resistors of 10ppm/k would give 0.35ppm/k, wire wound and metal foil have 2..5ppm/K max. T.C., which will yield 0.07... 0.18 ppm/K.

The metal foil resistors have no appreciable advantage over the wire wound types. Even the hermetical sealed VHP202Z, advertised as 2nd level standard and with typical T.C. of < 0.05 ppm/k will not give better results, because this extraordinary low value will not be achieved in reality. Also, the parabolic shape of the T.C. does not exist.

Picture 2 shows a high precision T.C. measurements on one of my 5 VHP202Z.
I measured values of 0.3, 0.6, 0.7, 0.8 and 1ppm/K, which is below the max. value of 2ppm/K, but far beyond the advertised typical 0.05ppm/K.

Therefore, an improvement can be made only by measuring the individual T.C. and selecting the ones with the lowest value. If R4 and R5 are matched concerning their individual T.C.s, this would improve the overall T.C. the most.

Conclusion: Using the LTZ1000, temperature set at 45°C, and using wire wound or metal foil resistors yield around 1ppm/yr. ageing. Using hermetically sealed, oil filled resistors may improve ageing to around 0.3ppm/yr.
Both resistor technologies yield a T.C. of < 0.2ppm/K.

Rem.: Sample variation of the LTZ1000 and other effects will lead to lesser stability figures.

(1) “ The Ultra Zener ... is it a portable replacement for the Weston cell?”,
P J Spreadbury, Meas. Sci. Technol. 1 (1990)
(2) “Setting new standards for DC Voltage Maintenance Systems. A Solid State DC Reference System”, John R Pickering, Metron Designs and Paul Roberts, Wavetek

[quarks]

Here is a short update, I received some more parts of my Vishay order, but still wait for the VHP101T. This feels like waiting forever.

Unfortunately (or better luckily) I am also quite busy right now, because I upgraded my lab with some gear that is planned to replace some/most of my actual calibration gear.   

@Mickle T., can you share where you found the Fluke 8508A details?
@Dr.Frank, are you planning to make a new PCB for your new ordered parts?

BTW Very interesting to see what is going on meanwhile.

@Dave + all interested, what do you think of the idea to try to design/make/build a kind of "EEVblog 10V Master Reference" based on LTZ1000 for this community?

bye
quarks

[Mickle T.]

@Mickle T., can you share where you found the Fluke 8508A details?

Thread about the voltage references: http://bbs.38hot.net/forum.php?mod=viewthread&tid=969&extra=page%3D1&page=12

[quarks]

Hello Mickle T.,

thanks a lot. I had a look also at some other topics there and found a lot of infos from you. I am really impressed, great work!!!.

Bye
quarks 

[branadic]

@Dave + all interested, what do you think of the idea to try to design/make/build a kind of "EEVblog 10V Master Reference" based on LTZ1000 for this community?

Still I believe 10V are very uncommon today, just usable to check your DVM/DMM in one range but nothing more as most circuits require a reference voltage of 5V or less.
Sure, it's all about stability, but wouldn't it be worth having a LTZ1000 based voltage source with all common voltages of todays need? On the other hand, wouldn't it be worth to have a decade voltage output to verify all voltage ranges on your DVM/DMM instead of only the 10V range? Mh...

BTW: I wonder if there is knowledge in here about that rejustor stuff by Microbridge? Some american guys in here taken their sample offer to verify the improved stability of standard references? How could the LTZ1000 be improved by such a value und tc settable resistor?

[babysitter]

Hi branadic,

Ich habe mir die rejustors mal im Zusammenhang mit Sensoren für den Aufbau von pga angesehen, ich denke trotz einstellbaren tc sind die letzten ppm Einstellung schwer. Und dass Prog-gerät plus Software eine über 1k Euro Sache.

[Andreas]

As the temperature stabilization of the LTZ is very sensitive, a high open loop gain may lead to instabilities. I'll see, just ordered 5 EA of LTZ1000, LTC1013 and LTC1052 from LT..

Hello Frank,

The instabilities can be handled by frequency compensation without spoiling behaviour at DC.
Why not the brand new LTC2057? I fear that the LTC1052 cannot deliver enough current if the short cirquit current is around 5mA. The LTC2057 is now available at DigiKey.

With best regards
 
Andreas

[branadic]

Thanks babysitter for that info. One grand for only the programming device plus software is inacceptable to give it a try and to come a cropper in worst case.

[Christe4nM]

BTW: I wonder if there is knowledge in here about that rejustor stuff by Microbridge? Some american guys in here taken their sample offer to verify the improved stability of standard references? How could the LTZ1000 be improved by such a value und tc settable resistor?

This is the basically the same question I asked in the Keithley 2015 teardown topic and was contemplating to ask here as well. Can anyone elaborate on this?

@babysitter, my German isn't that good that I can understand your post completely, but from what I understand you used those Rejustors and you found that even with "adjustable TC" it's hard to get the last ppm's down?

[Andreas]

Can anyone elaborate on this?

@babysitter, my German isn't that good that I can understand your post completely, but from what I understand you used those Rejustors and you found that even with "adjustable TC" it's hard to get the last ppm's down?

The key spec of precision is "long term stability".
On resistors we expect values below 25ppm/kHr or even less.

Mhm... What do you expect from a device which has a long term stability of 0.5%??? 50000ppm???
And when the adjustment range is +/-100 ppm I would not expect that you can get better than around 1/10th of it as stability. So in precision cirquits It might be useful at maximum to adjust the last 0.001% of trimming.

With best regards

Andreas

[Dr. Frank]

Still I believe 10V are very uncommon today, just usable to check your DVM/DMM in one range but nothing more as most circuits require a reference voltage of 5V or less.
Sure, it's all about stability, but wouldn't it be worth having a LTZ1000 based voltage source with all common voltages of todays need? On the other hand, wouldn't it be worth to have a decade voltage output to verify all voltage ranges on your DVM/DMM instead of only the 10V range? Mh...

BTW: I wonder if there is knowledge in here about that rejustor stuff by Microbridge? Some american guys in here taken their sample offer to verify the improved stability of standard references? How could the LTZ1000 be improved by such a value und tc settable resistor?

You're right.. for a complete calibration, you have to transfer the basic voltage reference (e.g. 10.000V) to the desired range voltages, i.e. 1kV, 100V, 10V, 1V, 100mV.

So you need a very precise and linear divider, like the 720A and the 752A, or the HP3458A.
They will aos deliver "uncommon" reference volatges as 5.0000V...
And you need a very stable voltage source, which can be compared against your basic reference by means of the dividers.

That's all. 

Frank