Calibratory D-105 DC Precision Voltage Reference Standard

[HighVoltage]

My most reliable instrument to measure this standard, is a calibrated Agilent 34410A (3 ppm accuracy in 1 year)

Once I get it, I will take some measurements and keep it running over a longer period.
With all our data of at least 2 units we should be able to determine if this system is stable.
I would not like to pre-judge it before we have any test data.

[Terabyte2007]

My most reliable instrument to measure this standard, is a calibrated Agilent 34410A (3 ppm accuracy in 1 year)

Once I get it, I will take some measurements and keep it running over a longer period.
With all our data of at least 2 units we should be able to determine if this system is stable.
I would not like to pre-judge it before we have any test data.

Cool, best I have in my own lab right now is the 34461A and 34401A, both in cal up to date.  Respectfully, don't you mean 30 ppm? 

[HighVoltage]

Respectfully, don't you mean 30 ppm? 

My finger was too fast, of course it is 30ppm (0.0030% +0.0005% in one year), thanks for correcting it.
Looking forward to your results on the 34461A and the 34401A
Will you use BenchVue to collect the data?

[Terabyte2007]

Respectfully, don't you mean 30 ppm? 

My finger was too fast, of course it is 30ppm (0.0030% +0.0005% in one year), thanks for correcting it.
Looking forward to your results on the 34461A and the 34401A
Will you use BenchVue to collect the data?

Absolutely!

[Dr. Frank]

Respectfully, don't you mean 30 ppm? 

My finger was too fast, of course it is 30ppm (0.0030% +0.0005% in one year), thanks for correcting it.
Looking forward to your results on the 34461A and the 34401A
Will you use BenchVue to collect the data?

The basic problem is, that they claim features / specs which hardly anybody can really check.
Herr in Germany, there are several volt nuts, who are able to really check stability , noise, and even uncertainty of this device.

I again offer to check at least the first two parameters.
Frank

[Smith]

I bet it's temperature stabilized, much simpler to implement than temerature compensation. I made one about 1,5 years ago just for fun. It was based on a Ref02 at 10V. Its kept at 45 degrees C with a temp sensor regulated heater in a small box. One year later it still said 10.00000V on my keithley 2000 ( after about 24hr warmup) like I had adjusted it a year before. By the way its adjustable by +/-100uV. Cant give any schematics because I later used this same schematic at work.

[Terabyte2007]

Respectfully, don't you mean 30 ppm? 

My finger was too fast, of course it is 30ppm (0.0030% +0.0005% in one year), thanks for correcting it.
Looking forward to your results on the 34461A and the 34401A
Will you use BenchVue to collect the data?

The basic problem is, that they claim features / specs which hardly anybody can really check.
Herr in Germany, there are several volt nuts, who are able to really check stability , noise, and even uncertainty of this device.

I again offer to check at least the first two parameters.
Frank

Yes, you are correct! That's assuming they are true to their claims. Then my best unit will be down in the noise. If that's the case, then it would be respectable. How would you propose testing the first two parameters?

[Terabyte2007]

I bet it's temperature stabilized, much simpler to implement than temerature compensation. I made one about 1,5 years ago just for fun. It was based on a Ref02 at 10V. Its kept at 45 degrees C with a temp sensor regulated heater in a small box. One year later it still said 10.00000V on my keithley 2000 ( after about 24hr warmup) like I had adjusted it a year before. By the way its adjustable by +/-100uV. Cant give any schematics because I later used this same schematic at work.

That should be easy enough to tell once I get it. Too bad you can't share! 

[Smith]

Its not that hard to design. Just some basic design, a good reference ic and some testing to eliminate ac voltage. Thermal coupling is simple too with a small piece of copper and thermal paste

[HighVoltage]

I again offer to check at least the first two parameters.
Frank

Hello Frank,
How about I will be sending it to you after I have performed a few basic test's.
Would be great to get your report on it.
Thanks for this offer.

[Dr. Frank]

I again offer to check at least the first two parameters.
Frank

Hello Frank,
How about I will be sending it to you after I have performed a few basic test's.
Would be great to get your report on it.
Thanks for this offer.

Yep. We should have much
 more ring comparisons,here in good ol'...

[Terabyte2007]

Ok, this is the preliminary. It arrived earlier than I was expecting so I did some quick testing using my 34401A. I will do much more in-depth testing in the days to come with my 34461A and BenchVue. I have limited time tonight to do anything further but was excited to give a quick glimpse of the preliminary results.

I allowed a 1-hour warm-up of both the D105-10 and my 34401A. Input 14.5V DC from my best PSU and was getting a stable 10.00000 VDC with the occasional flicker to 10.00001 VDC. The unit came with a small Certificate of Calibration with the following:

Date: 01/31/2015
Time: 10:00:00 UTC
Make: Calibratory
M/N: D105-10 DC
ID: 22** (Edit for Privacy)

Environmental:

Temperature: 68F/19C
Relative Humidity: 37%

Test Method:

Null offset to Fluke 732A Precision Voltage Reference

Parameters:

Input Voltage: 14.5VDC
Output Voltage: 10.00000.57 VDC
Actual maximum error: <=5.7 uVDC - 0.6ppm
Specified maximum error: <=20 uVDC, 2ppm

Signed with initials and dated 1-31-15.

I have included some pictures of the setup. There will be more to come, stay tuned!

[Lightages]

So far that is impressive. I guess we wait to see how long it stays that way.

[Lightages]

I was thinking of buying one, but many sellers on ebay now are sing a silly door to door shipping system that almost doubles the price. Too bad.

P.S. I just sent them this message:
"I was interested in buying one of these from you, but your shipping system kills the deal. I don't know why many people are opting for this "easy" shipping system but in my case your only option listed adds $60 for shipping and importation. I am Canadian living in Chile and you should know that your price for importation will not be valid in Chile and most of South America. The country will decided to make some people pay the taxes on the shipping, plus the charge you have taken for the importation fees, and then ask for their calculated importation fees on top of that again. Please offer a real option like just normal US post with tracking. This new form I see on many sellers offerings makes it impossible to buy things and kills your sales.

Can you offer me a simple shipping method instead that does not cost me twice the value or more in the end?"

[rx8pilot]

$60 is not bad for 'intl shipping. I have to ship small things around the world about 10 times per week and it's just a pain in addition to the expense of the services. From the USA to just about anywhere, a small box is better than $45 plus the PITA paperwork for postal service which is weak, but workable. FedEx and UPS are way more expensive even if you are only shipping a paper clip. There is no relationship between parcel value and shipping cost with the exception of very high value insured packages.

The unit is only $100 so I would not care about an extra $60 for s/h.

[Dr. Frank]

Hi,
for US => Germany , they charge 13$ for postage only, and estimate another 25$ for tax and customs, which is correct.

I recently exchanged a small parcel with a volt-nuts in Seattle, and it was about 20$ each way with USPS and DHL, and vv., but w/o tax/customs fee.

Therefore, I don't understand these 60$ for door to door?
If tax and customs fee is necessary, I have to personally pick up the parcel at a customs office, without additional fee than these mentioned 25$.

So you may specify for your country also, that the customs handling is not to be done by an additional service provider, who may charge extra money.

Frank

[Dr. Frank]

..
I allowed a 1-hour warm-up of both the D105-10 and my 34401A. Input 14.5V DC from my best PSU and was getting a stable 10.00000 VDC with the occasional flicker to 10.00001 VDC.
..

Test Method: Null offset to Fluke 732A Precision Voltage Reference
...

As you are not actually showing the Null Offset method against your 732A, I assume, you keep your 34401A always calibrated to the 732A, instead?
And I also assume, that your 732A is calibrated in a traceable manner?

Well, then this D-105 device really would be very good for initial absolute uncertainty.
But maybe it would be a waste, if it drifts too much over time.

And in this sense it will be interesting, how the ordered part will survive the much longer journey (thousands of miles and weeks of time) to Germany..

Frank

[babysitter]

US to Germany:  <22EUR = 0%.  22-120EUR= 19% EUSt tax. >120EUR=19% tax plus customs.
Consider parts without a CE mark to be more complicated - to the point of impossible for electronic devices.
(If it is just a component which cannot be used standalone, not a device, it might work without CE.
As a ham you can import without CE and declare yourself that you take care for conformity.)

[Dr. Frank]

US => Germany is 0$ for goods up to 22 EUR(), 19% EUSt for goods between this level and 120 EUR and above that, tax kicks in. Thresholds might vary over time.

You mean 19% tax  plus customs fee, or toll (German: Zoll) kicks in above 120€?

Frank

[quarks]

..
Test Method: Null offset to Fluke 732A Precision Voltage Reference
...

As I read it that is the setup from the seller and looks fair to me.
Because a real traceable calibration certificate would probably cost way more than the offered reference itself.

[saturation]

Wow, this is very encouraging news  .  The unit reads to its factory calibration after marked environmental stress while powered off  .

At this time, peak winter in the US, even if calibrated at room temperature, packages are stored in unheated mailing areas and shocked by manhandling, even if packaged well.  We could also add the possible shock of airline flight to 0.5 bar and uncontrolled humidity exposure.   Unlike late spring and early fall, when room temperature is nearrt ambient, high winter and summer puts a large thermal difference on references shipped in uncontrolled environments.

Dr. Frank's points are healthy skeptical critiques that need further testing, but its looking good so far.

Quarks comments are too, but we cannot expect standard metrology practice from this device or maker, given its pedigree.  Should it survive Terabytes further scrutiny it would be good to see Dr. Frank's testing or send it to an NVLAP metrology lab for an incontestable analysis.

Ok, this is the preliminary. It arrived earlier than I was expecting so I did some quick testing using my 34401A. I will do much more in-depth testing in the days to come with my 34461A and BenchVue. I have limited time tonight to do anything further but was excited to give a quick glimpse of the preliminary results.

I allowed a 1-hour warm-up of both the D105-10 and my 34401A. Input 14.5V DC from my best PSU and was getting a stable 10.00000 VDC with the occasional flicker to 10.00001 VDC.

[TunerSandwich]

Personally, I doubt the performance , once from the capabilities of these refs, 2nd because the description is too enthusiastic, and there's a strange list of compensation features... Like these guys, which claimed supercaps from nanoparticles, and Free Energy, and so on.

Hello,

I also doubt that this performance (+/- 2ppm) is possible in a plastic package.
The better value of long term stability (5ppm/kHr) is usually only for the metal can package (TO-99) which is now marked as obsolete. For the plastic package 20 ppm/kHr are specced. The "typical" 5ppm/kHr of the foot note after 168hrs are usually only valid for lab conditions with constant humidity. In real life applications the plastic packages will drift with humidity.

Further all accuracy specs in the datasheet are additive. So a good performance is only possible under very stable conditions.
If you let the power supply at +/-2V accuracy the +/-2 ppm are already used up by the PSRR of 1ppm/V. And that parameter is meant without self heating due to changing power consumption.
If you really draw 10mA the 10ppm/mA will sum up to 100ppm. (+additional drift due to self heating).

"No soldering" (= socket for the REF102) sounds good for the first. But over time when contact oxidation takes place you might get strange effects by a socket.

There is no statement about pre-aging for the D-105 DC.
In my opinion at least 1 kHr continuous operation would be necessary before first calibration.

So all in all:
Although the REF102 is a excellent reference (in TO-99 package)
in the plastic package I would not expect to get better than 10-20ppm over 1 year with following measures.
- well stabilized power supply around 15V.
- no load on the reference output
- temperature controlled operation

A better choice for a reference would be the AD587 in CERDIP package.

In my ageing experiment I am running one REF102AP against several AD586LQ (5V) references.
14V stabilized power. 50 deg C stabilized temperature. No Load at reference.

Within first 70 days ageing drift was 10 ppm and seems now to stabilize.
When regarding stability the plastic package is far not so stable than the CERDIP devices.

With best regards

Andreas

That isn't entirely correct.....the REF102C is rated at 5ppm/1000hrs AFTER 168 hr stabilization period....

see note #2

also here is a list of actual units tested in Ti lab report.....if you hand select it's not so cut and dry....(note that the lab report is on the CM package...which is now "obsolete")

I have used the REF102C with excellent results...in my own ref designs....it's one of my favorite inexpensive ref IC

[Andreas]

Hello,

I bet it's temperature stabilized, much simpler to implement than temerature compensation.

with the power consumption of less than 2 mA there will be no heater:

https://www.eevblog.com/forum/testgear/calibratory-d-105-dc-precision-voltage-reference-standard/?action=dlattach;attach=133947

so most probably the compensation is similar to the Geller  SVR-T with AD587LQ.
(which I think is the better choice for long term stability).
http://www.gellerlabs.com/Voltage_References/SVRTH/SVR%20THR%20Sch.jpg

So don´t get me wrong: the d-105 dc reference standard is a good price for the service of importing the volt into your home lab. But I would never rely on long term stability.

From my friend volt nut I have the indication of around 0.3-0.4 ppm/% rH for the REF102 in plastic package.
Time constant of epoxy packages with humidity is around 3-7 days reaction time.

So my suggestion would be to do verification measurements in february / march when there is low humidity and in august / september with high humidity.

also here is a list of actual units tested in Ti lab report.....if you hand select it's not so cut and dry....(note that the lab report is on the CM package...which is now "obsolete")

So they have no problems with humidity with the CM package.

From a 5 V Reference (LT1027CN, 2 samples) I get a humidity coefficient of around 0.5ppm/%rH
see attachment where you can see that the uncompensated curve follows a 5 day filtered humidity curve.
X-axis is in days so you can see the seasonal changes of humidity. The ageing of these well aged devices is less than 2 ppm/year in average but humidity excursions are around 10 ppm within 1 year.

With best regards

Andreas

[babysitter]

Add humidity sensor, inject filtered compensation signal into zener current ?

[Andreas]

Hello babysitter,

Even if I tried this the resulting error cannot be better than
around 2 ppm due to humidity as you can see in the simulated corrected curve.
(I had also the idea of compensation and already a SHT15 sensor).

But calibration time would eat up the rest of my life.
For 1 device I would need at least 2-3 years. So I abandoned it.

With best regards

Andreas