Ultra Precision Voltage Reference and Maker Faire Hannover 2018 

That's what Electrical Engineers do. Lol.

My main fields are measured in "does it run or crash" (computers) and "is the surface smooth enough to actually resemble what I printed" (3d printing purely functional parts). The general population thinks computers are the ultimate in precision electronics, but in reality, they are the ultimate in engineering around the lack of precission. It's incredible how in a world where we are capable of precision electronics engineering ranging in the nano ranges for all practical measurements, computers are now at a point they are verging on being considered non-deterministic because of all the humbuggery to keep them running with excessive heat, the dirtiest and wildly varying frequencies, dirtiest and wildly varying voltages and unreliable digital data. And this done on gate feature sizes on the single digit nano scale sending signals at up to and beyond 200/picoseconds, in parallel, through billions of transistors. To the point the reliability of the data that comes out of them is questionable enough to become a real problem in our data driven world. People are developing quantum computers independently, but the further downscaling of binary computers may just accidentally result in our compensation efforts for the phenomena at those sizes to bring forth quantum computer capabilities and features based in standard binary electronics.

Anvilshock explain

Correct name is 'P...-T... Bundesanstalt', meaning 'Federal Institute for Physics and Technology'. This institute actually exists. 'Badeanstalt' (this also exists) sounds and looks remotely similar but translates to 'bathing institution' and is what public swimming pools are called in typical German bureaucratic jargon. The joke is in the intentional malapropism, a trope quite common in German humour. It may be either purely decorative, to sprinkle some flavour into one's speech, it may be a jab at the original term's jargonesqueness itself or at some aspect of the thing for which the term is the name. An (US-) English equivalent might be 'Federal Bureau of Infestation'. The intentional malapropism here may again be entirely nonsensical and unrelated or may carry an ironic meaning indeed. Conceivably, the FBI example could lend itself to alleging a practice of 'infesting' a place with 'bugs', ie. placing listening devices, but I could neither confirm nor deny any allegations as to the actual FBI's methods and resources without the Secretary's approval; after all, I'm not associated with the FBI, and certainly not with the IMF, either, and even if I were, I would not be at liberty to disclose so. But since I'm not, I couldn't, even if I wanted to. Not that I'd want if I were. Ahem.

@@Anvilshock thanks for the explanation. Bob dabiuld look up 'german humor' on Wikipedia. It's very good. I think german humor has something to to with german grammar and german culture; but I suppose you could say that about anything.

@@superdupergrover9857 Thanks for the nudge; good stuff, as you say: en.wikipedia.org/wiki/German_humour

@@Anvilshock That was the most elegant explanation for a joke I've ever read.

i once got free sponsor from PCBWay, you can have a try

@@RealElaineee What do you mean?

components are cheap. next to free when you recycle.

pepsi jazz, check out some of Bob Pease articles on just how sensitive some measurements can be.

@@anullhandle but couldn't this problem be solved with a faraday cage?

@@Ramog1000 , Down the rabbit hole past faraday cages, exposed copper gold plated traces, multi layer guard rings comes making connections for sensitive nodes in air off the board, minimizing the air space to reduce the effects of cosmic rays, thermal isolation, acoustic isolation and other analog voodoo.

He should have used a lead enclosure to protect it from cosmic particles.

Well that's probably on me, I approached the guy with "Excuse me, can I have a quick voltage measurement?" instead of "Make way, here comes the best thing you'll see all day!"

Note for future : have your box with proper cable ends termination to copper spade lugs. So you can connect to proper meter in no time and get expected sdev

Awesome! I've been thinking about a box from soldered mu-metal and hermetic feedthrough-capacitors for the other board, but I bet there'll problems with customs when shipping hermetic containers?

It seems that both the spehere and the watt balance will be used. They want the Avogadro constant and the Planck number to be more accurate it seems. We will know more on 16 Nov 2018 though when the conference is over.

Were the components you used only available on request?

For easier schmearing.

Vely intelesting... ;)

cutting them afterwards let the joint vulnerable to oxidation (microcracks)

if thats your concern then you can always touch it to reflow solder over any exposed copper (if your not lacquering it in the first place), where thats still preferable to actual damage from thermal shock/ overheating

but wouldn't reflow will introduce way more heat ? I mean the DIY reflow oven/plate method. The only thing i know which is superb is that zinc wave bath :D.

Personally I just crystallized tin and lead directly on the pads from solution so the parts weren’t overheated on my voltage reference.

haha this is a joke , right? :) if not video pls

In the same room was a guy with a tesla coil and some HAM radio operators (like me *hrhr*) broadcasting at 50W... So yeah, all those measurements were a bit problematic, but I guess this equipment is well shielded.

Yes I see the Uni-T dmm in the corner that they used to do the fresh calibration 🤣

Try Germany.

Why is the constant power important for a voltage reference? It can't just be temperature related can it?

Console It's to do with the subatomic dielectric permeability coefficient of non-linear electron transfer substrates. The longer you leave the reference turned on the less you have to factor it in.

He got a special delivery from Unobtanium Planet.

Given the environment, I would bet a good portion of that deviation wasn't even his fault, but the fault of all the electrical noise at the fair.

I can top that, 3€ from a store called Action :D

How about my 1$ meter xD

There was a really nice weston-cell-based voltage standard once, 6 cells in a big, multi-walled metal oven that's kept at 30 +/- 0.1°C ....... who knows, maybe I can get my hands on one :)

Maybe if he could beat those things 100 times a second...

nickt cause if you go through the trouble of fine smd components why not also shrink the board ? I mean i only etch 2 layers thats why my boards are larger but if u order them i would try to get the lest space possible

I don't think it is about shrinking the board. The additional layers are most likely used for decoupling or an otherwise better layout in terms of heat isolation(since they want the least amount of copper in the middle of the board possible) or interference for example. I very strongly believe that the 4 layers were a performance consideration more than anything.

LMAO

Sexiest voice on youtube

It could be more like 25 people.

Me too

It is the most stable voltage source one can have for ~300 USD. I can compare my other devices to it and be reasonably confident in their drift. I could also use it as a reference inside another device, who knows :)

Thanks for the reminder; I only discovered this term a couple weeks back, and immediately forgot it again :-) en.wikipedia.org/wiki/Bob_Widlar

Voltage is always relative. There is no absolute voltage. Voltage is just the difference in potential between two points. 0V is whatever you define as 0V. Thats why multimeters have two leads. You define the ground lead as 0 volts and then calculate the difference between the two. Its not actually objectively 0 volts. What you have to find out is the difference between two values, rather than an absolute output value. But yes, you have to measure that difference somehow. Its only as precise as you can calibrate it. Thats the case for any sort of voltage reference though.

Not necessarily. Even without knowing it’s voltage, the fact that it changes so little over time is useful in of itself (if you measure it with a moderate cost multimeter, you are actually measuring the drift of you meter, not the reference)

@@reps beg for them

Lol what

I googled it and this is what i found: Applications • Precision Data Acquisition Systems • PLC Analog I/O Modules • Field Transmitters • Portable, Battery - Operated Equipment • Industrial Instrumentation • Test Equipment • Power Monitoring • LCR Meters Still not very clear to me. Would I build a Multimeter with this?

Yes, every multimeter needs one, but only the top high end models will use an LTZ1000

Es soll eine spannungsreferenz sein, nichts weiter. Lediglich eine bekannte spannung die so konstant wie möglich bleibt damit man sie mit anderen vergleichen kann. Die kann natürlich zum kalibrieren von voltmetern benutzt werden, könnte aber natürlich auch direkt in einem voltmeter verbaut sein, die haben ja auch intern eine referenz.

danke... interessantes thema, hab ich mich so noch nie damit auseinandergesetzt. immerwieder erstaunlich, was es nicht alles gibt - man nimmt das immer als selbstverständlich, dass beim netzteil die angegebenen spannungen rauskommen. das gerät was marco da gebaut hat, erweckt mir aber dann doch eher den anschein, als ob es mehr mit schwarzer magie zu tun hat als mit, sagen wir, wissenschaft. allerdings würde mir jetzt spontan auch nix einfallen wie man eine konstante spannung erzeugen könnte.

Naja, das ist in dem fall eine zener diode. Das bedeutet ab einer bestimmten spannung lässt sie strom rückwärts durch(und das sind in dem fall wo die 7.1V herkommen). Eigentlich tun das alle dioden bei irgendeiner spannung, aber zener dioden sind halt dadrauf ausgelegt dann nicht kaputt zu gehen.