EU - CalClub

[e61_phil]

Here in germany (and GB as well) is the standard cal temperature 23°C and not 20.

[BFX]

Please write me down to the list 
LTZ rulez 

[hwj-d]

Short conclusion after the Stuttgart summit regarding a concrete implementation of the EU-Calclub.

My request to include the EU-Calclub design and organisation question on the speaker's agenda in the short term was rejected for understandable reasons. It simply didn't fit in in terms of time and topic. Dr.Frank was kind enough to integrate the topic into his remarks by once again clarifying the tempco problem of shipping at winter temperature.

But that means nothing else than that we, as interested parties of an EU-Calclub, did not get any further with the organizational question and the reflection to a technically feasible variant as well as with the clarification of most different requirements to a reference than it was already addressed here in the forum anyway.

The question of preventing hypothermia does not only arise in this direction, however, if the possibility of overheating in summertime can be given on an equal footing. Preventing subcooling by means of polystyrene or other materials obviously does not work because it merely avoids emitting too much heat in too short a time, but not being able to heat passive elements additionally. And on the other hand we don't want to build perhaps
self-igniting paraffin-filled M...-Cocktails.

So my conclusion for the time being is that we have to get away from the idea within the EU Calclub of being able to build a highly qualified reference in our own rooms, but then also want to send it on its journey with this claim. For the same reason, it is likely more important to provide a higher quality reference to control a 5.5 - 6.5 digit DMM at the receiver, than to test the reference as an end in itself for the transmitter. Which doesn't rule out that the reference will be calibrated more and more precisely, depending on which qualified hands it passes through.

In my opinion, it is even completely irrelevant whether a "usage reference" as such(!) has suddenly migrated 5 ppm after the star-shaped control, if this is referenced accordingly in the cloud. Yes, it should even be expected that this will happen all the time!

I will probably first choose two of my LTZ1000 in a few days, and then send them to the first interested people under these conditions, in order to concretize our request further. In the end, it is of course up to each sender to decide under which circumstances he/she wants to make his/her references available.

I would like to thank every active participant, designer, sender or receiver, as well as all necessary constructive critics.

Regards,
H.-W.
 

[branadic]

My conclusion is:
A star comparison (send it to a receiver and back to the transmitter) is better then a ring comparison. This way you can see, if there was any change during transport. Furthermore the temperature of a LTZ1000 based reference should never fall below -15°C/-18°C.
Another point I extracted for myself is that the reference should be temperature compensated to zero t.c. by R9 and that reconstruction of the output voltage is possible as done in Wavetek/Metron/Fluke 7000 after a power down. Thus cool shipping of a reference with multiple styrofoam boxes stacked together should be possible.

-branadic-

[e61_phil]

Why not shipping hot? My LTZ1000 "Brotkasten" runs around 5 days from batteries.

Is shipping cost really an argument? It is still below 10€ even with heavy lead batteries and a protective case.

[dietert1]

At the metrology meeting Stuttgart i thought i wanted as transfer standard a 10 V reference with builtin microvoltmeter, so that the transfer is a direct comparison and independent of another reference in a high resolution DVM. A 12 Bit AD converter with a range of +/- 100 uV would be good enough for this.

This could be something like a MSP430 or Cortex M0 that runs on a small battery. At the same time i'd like to replace the analog thermostat of my AD587 reference by a digital one including temperature logger. The current analog implementation lacks monitoring during travel and voltage transfer. It stays "hot" for about a week on a usual USB power bank.

Then it's a minor step to implement a "freeze protection" mode where the thermostat is off and will turn on only to avoid temperatures below 0 °C. With this concept the whole device including the battery should travel safe for one or two days and the transfer should be much easier, too. So the standard returns faster.

Don't know when i will have time to work this out.

Regards, Dieter

[cat87]

I think within the western part of Europe,  sending the reference powered on is doable,  both in terms of shipping costs and time. Within the 5 days that e61_phil mentioned is doable. However,   for the eastern part,  I think it takes considerable longer than 5 days.

 On the other hand,  shipping the reference cold means that each recipient will have to have the same temp in the 'lab' as the one at calibration time....I personally can't do this and temperature deltas as high as 5 degrees are not uncommon.

All in all,  probably sacrificing resolution and low noise so that the reference is available to a larger part of members in Europe would be ideal...
Or maybe even have 2 references running around?  (one hot and another cold) so that each can pick and choose depending on the estimated travel time

[3roomlab]

just to add some data to the "shipping problem"
there are many articles about shipping temperature on/in US continent
I cant find much on/in EU
but this 1 article, not really EU but "japan to netherlands". the netherlands portion temp wiggle is not small.
https://interdry.wordpress.com/2012/01/27/ocean-container-temperature-and-humidity-study-2/
the span is nearly 40C. I think this could make permenent shifts in ppm (as observed by DrFrank in his findings)
not sure how they did the measurements, but looking at some US based transport papers, they also have similar wiggles.

middle of year


year end


this 1 is from some kind of wine transport log


I put this into LTspice, assume some thermal mass numbers which is approximately familiar. this is a 0 to 50C swing going thru box/VREF insulation stepping of 10 to 400 C/W. the timescale is 1:1.
insulation have some effect esp at 400C/W. but 400 C/W means (for a 5x5x5 cm box) something like 8inch of 0.035W/mk foam, item becomes an elephant! doubling the surface area of the box = double the thickness 16inch. size becomes a factor? maybe?
another rough simulated figure @ 400C/W approx saving in internal heated power is 4x, @ 200C/W saving in internal heated power is 2x. either a crap ton of insulation or crap ton of battery I suppose there is no escape from power of mother nature. (but these are all LTspice based numbers, and I think the temp ramp is too slow, should be nearer to trapezoidal square)
edit1 : I think if the shipping is long and far, the insulation could be thick. if the shipping is short and near, the battery could be heavy instead. but Im not sure

[Echo88]

Can you show us the Brotkasten?

[hwj-d]

I can't mention enough:

• We have more than one senders with more than one reference. For the very first beginning I spend two ltz1000 maybe my orange dekabox too.
• Others may have better Refs, bigger styropor packages, hot or cold ones.
  That all belongs to the senders convenience and expediency.
• The shipping pattern was star-shaped from the very beginning.
• It should announce itself senders, which can make now (this month) a ready reference available.
• Ideas, building plans, further ideas are of course welcome. Once they have been assembled and tested, they can also be securely integrated in the future, but we should concentrate on fully assembled references for the current project.
  What is available now?!
• The only condition is an online maintained list of the actually available references, their current state together with their whereabouts, and a current measurement protocol depending on the metrological capabilities of the receiver and transmitter.

My concluson (and proposal) is:
 do what you can do in the real moment, but
   keep it simple!

 

Thanks,
H.-W.

[wim.de.jong.59]

Please, put me also on the list to start with.

I've the following equipment:
Fluke 5450A
Fluke 732A, 2*
Fluke 731B, 2*
Fluke 742A 10K
Agilent 3458A
Agilent 34401A
Fluke 8505A
Fluke 510A, 1KHz

Maybe, its an idea to start with a Fluke 731B which is characterized for 20-25 C?

Wim.

[The Soulman]

Please, put me also on the list to start with.

I've the following equipment:
Fluke 5450A
Fluke 732A, 2*
Fluke 731B, 2*
Fluke 742A 10K
Agilent 3458A
Agilent 34401A
Fluke 8505A
Fluke 510A, 1KHz

Maybe, its an idea to start with a Fluke 731B which is characterized for 20-25 C?

Wim.

 

Welcome to the forum Wim,

Are you a hobbyist, or do you actual need all this gear? 

[Storm]

Please also put me on the list as well...

Cal Equipment:
Fluke 5440B (2x)
Fluke 5101A
Fluke 5200A (to be restored with one spare unit for parts)
Fluke 731A
Fluke 720A
Fluke 845AR
Fluke 845AB
Fluke 8502A
Fluke 8505A (2x)

Next project is a LTZ1000 REF-Standard to calibrate all of the above. I will use Dr. Frank PCB, new LTZ1000ACH and precision Resistors from Edwin Pettis. Still need to read a lot about this topic but the metrology virus already struck me hard...

I am a hobbyist but intend to use this equipment to restore and calibrate all my other stuff!

[Zucca]

As soon I got the 10V ref. Board form TiN, I would be happy to join the club with my other expensive toys.

[alex-sh]

Are we getting anywhere with EU Club?


Thanks
Alex

[Martin.M]

is the club only open for people who own expensive toys or also for some who have anything to check up?

Keithley Calibrator / Source 263  needs cal (Triax) so I will not offer to use it for anything..

I have a Fluke 883A what is a full year ago gone to a friend from fluke germany, but I think he will not find time to cal + adjust my oldie, so 883A needs a cal / adjust also.

Here is a restorated Tek 213, the intern multimeter is to cal 

Also a restorated Fluke 8500A DMM, to cal

.. and some other oldies .. on the makerfaire Hannover it was not possible bec. the cal place from the PTB was always busy.

Martin

[dl2ocb]

Hello to all Voltnuts,

a little nice CalClub would be fine.

I got some gear that should be compared to get better Feeling about Stability and Accuracy.

GenRad 1440 10kOhm and 100kOhm last Cal in 1968
2x LTZ1000 7V KX References (build 08/2019)
2x Fluke 720a KVD
2x HP3457a DMM
Russian Decade Resistor

I would be happy to contribute with Measurements in RF Domain (partly to 40 GHz). (https://dl2ocb.de)

Hope to meet some of you in Kassel to the DARC Funk.Tag 2020, im there rocking the Funkmessplatz (https://funkmessplatz.info)


CU
Stefan

[branadic]

After one year of the initial idea to the EU or at least DE cal club, how are things going?
Unfortunately Metrology Meeting 2020 had to be canceled, so chances are gone to compare to Michels Metron/Wavetek 7000 (M69) and even more, I didn't had the chance of comparing references to M71, that had been circulating last year and also showed up on Maker Fair

So since I have a few things that I would like to compare to some well known 10.xxxxxV reference, I'd like to ask: who can contribute something to a star case cross comparison? Current temperatures outside are good and in case we are still concerned about temperature with power off, I'm sure I will find some proper styrofoam box at work.

If noone has a proper reference, I can contribute a temperature compensated LTZ 7.xV with AC wall adapter and cable for lab power supply. I can measure it against our 3458A and K2002 but also my R6581D and S7081 prior shipping. By star case comparison and reporting the measured value to all contributors, we can recalculate our own readings. Maybe we find some contributors with calibrated gear and cal reports, that could be added each time the reference is shipped further?

Awaiting your response and hoping, that it won't take another 120 days or year to get something up circulating.

-branadic-

[The Soulman]

I can put together something small with a selected (1ppm/dgr c) LT1021dmh-10 and a non-characterized 10KOhm
ESI wirewound resistor, that could and should be done after assembly, also the LT1021 needs multiple thermal cycles after
soldering to make sure it has not shifted or is prone to shifting.
After that at least two weeks of monitoring, also can not give a reading to any meaningful calibration uncertainty, I'd
like to think I'm in the 10ppm ballpark, but more likely 40ppm or so.
So preferably first receiver should put more accurate numbers on it.

If anybody sees use for such let me know and I put it together, although already I'm aware it may not please the 3458/732 crowd.

[Dr. Frank]

Hello André,

as always, I offer my transportable LTZ and a 10k references, aligned during last Metrology Meeting (1y.) and to M71 (1/2 year ago). Plus a better attempt for drift prediction. Plus styrofaom box to transfer them safely.

I am currently updating LTZ #5 on battery operation, i.e. AnJa's design. Very stable reference, still, after all that Change stress.
Doubt that it can be sent via DHL, due to mechanical constraints. Wonder if AnJa ever shipped any of his references.

I'd prefer a small scale volt-nuts meeting in Stuttgart region, instead.
I've got time, working in Home Office, and I would contribute local Äppelwoi, and Handkäs with music, if not declined.

Frank

Edit: There's really an English Wiki: https://en.wikipedia.org/wiki/Handk%C3%A4se

[branadic]

Hello Frank,

the idea of a small scale meeting is very much appreciated. 

Unfortunately the battery don't last long enough to ship the references based on Andreas's design, as shiping is somewhat unpredictable in time, but it is a good reference for personal transport.

-branadic-

[Andreas]

Hello,

I already thought of a power bank (there are versions up to 50000 mAh available) and a small step up regulator like in my AD587LW reference to buffer the internal batteries during land based transport.

with best regards

Andreas

[MiDi]

Is there any information on how much hysteresis can be expected from LTZ1000A (e.g. setpoint ~50°C (12.5k/1k)) when switched off several days (room temp ~21°C)?
If this is negligible, would it be feasible to add an efficient circuit to power only the heater and maintain e.g. min 21°C inside LTZ to achieve significant lower power draw for shipment [at low temperatures] and get rid of significant hysteresis?

Problem with powerbanks is that there are restrictions for shipment, in particular if they are switched on.
If I remember correctly all Li based cells are problematic, but lead acid, NiCd, NiMH do not have significant restrictions - even when delivering power.

[mendip_discovery]

Here in germany (and GB as well) is the standard cal temperature 23°C and not 20.

This is general aim for calibration labs in the UK (UKAS etc).
Mech Labs it is 20°C ±2°C
Electrical Labs it is 22°C ±2°C

The difference I am lead to believe is because of the US manufacturer of meters and a preference for slightly warmer conditions.
The closer your temp is to 20 or 22 can help reduce your uncertinty.

[MiDi]

Thanks to Dr. Frank I got a 10kΩ and 7V Reference to compare against my ancient HP3458A and to do a transfer to my references.
I have essentially no experience with these kinds of calculations, please forbear with and correct me if I made a mistake.

The specs of HPAK 3458A and the given uncertainties of the references from Dr. Frank are expected with an extended measurement uncertainty (k=2).
For the reference values I used the mean values provided by Dr. Frank before and after return and added half of that drift to the uncertainty:

LTZ #2 (powered from delivered wall-wart):
7.118223V + 0.5ppm/2 = 7.118 224 8V
EMU: 2ppm + 0.5ppm/2 = 2.25ppm

My results (Mean from 16 measurements with 100NPLC, AZ ON):

7V:	7.118 241 0
-7V:	-7.118 237 3
Offset:	1.52µV
StDev (7V):	206nV
7V offset corrected:	7.118 239 4
-7V offset corrected:	-7.118 238 8
Value Offset corrected:	7.118 239 1
T-EMF & INL 3458A:	320nV -> ~0.05ppm of measurement
ppm diff:	2.0

Result 10V range: +2ppm +-2.25ppm

V336 10k:
9999.979Ω - 1.4ppm  + 0.5ppm/2 = 9999.968Ω (-1.4ppm is extrapolated drift from last cal by Dr. Frank)
EMU: 3ppm + 0.5ppm/2 = 3.25ppm

My results (Mean from 6439 measurements ~15h, APER 1, DELAY 1, OCOMP ON, AZ ON):

10k:	9999.963Ω
StDev:	0.28ppm (Mean of moving average from 16 measurements)
ppm diff:	-0.4

Result 10kΩ range: -0.4ppm +-3.25ppm

I am impressed by the results - they are much better than I expected.
The drift in last 4 years compared to 2016 are in the order of +1ppm/year for 10V and -1ppm/year for 10kΩ range.

The T-EMF & INL of 320nV are around 0.05ppm of measurement (7V), that is in the order of the spec for INL of 3458A.

Attached the results from last cal of pre-owner 2016, from Maker Faire 2019 and last but not least current results from 2020.