Bottled! LT1021CCN8-10

[zhtoor]

-

[Andreas]

Humidity is relative easy:

just record %rH and measure your reference output.
Minimum over one %rH cycle (around 9kHrs)
Then you will see some correlation according to the seasonal changes of your local %rH.

Of course the reaction time to %rH is not very fast so you will have to filter the %rH somewhat
(perhaps 5days time constant without the rosin with rosin somewhat more).   

By the way: you should work on your resolution. I measured around 0.5ppm/%rH on my 2 LT1027 devices.
(perhaps measure in the 100mV range against another stable reference)
And of course you have to calculate out T.C. and ageing before you can get the humidity coefficient.

with best regards

Andreas

[cellularmitosis]

You can use saturated salt solutions to create regulated humidity environments.  My plan is to start testing resistors, vrefs, etc, by simply swapping the lids of these mason jars after ~1 month, to give them a step-change in humidity.

as soon as I get around to it... 

https://www.eevblog.com/forum/metrology/t-c-measurements-on-precision-resistors/msg1481424/#msg1481424

https://www.eevblog.com/forum/metrology/calibrating-humidity-sensors/

[Andreas]

I do not like salts near my precision gear. 

[cellularmitosis]

These would likely have to be sacrificial units to determine the humidity coefficient

[Andreas]

 

hope you did switch off the WIFI of your laptop (near the picotest) during measurement.

By the way: a Allan deviation diagram of the picotest against a 34401A on a stable source (heated LM399) would be interesting.

With best regards

Andreas

[Andreas]

Hello,

for Allan deviation I would prefer a long term stable source.
Short term (noise) is secondary in this case.
It will only lead to a somewhat increased starting point of the Tau-sequence.
A NiCd battery may drift over time/temperature more than the instruments.

with best regards

Andreas

[David Hess]

What about mounting the reference inside a hermetically sealed environment and drawing a vacuum?  Then there should be no effects from barometric pressure or humidity.

[branadic]

What about mounting the reference inside a hermetically sealed environment and drawing a vacuum?  Then there should be no effects from barometric pressure or humidity.

Makes only sense, if you have a ceramic board instead of FR4 and none outgassing components mounted. The better way is to fill the inside of your sealed environment with some inert gas like nitrogen or argon.

-branadic-

[branadic]

Use a (steel) case with feedthroughs, put your board inside, fill it with oil, wax or fluorpolymere. Use a soft-annealed copper seal between case and lid. Don't use bare forging alloys as they are none-hermetic.
You can use electro-plating to cover your steel case with Sn (tin) for soldering the feedthroughs. Outherwise you can use brass or aluminium and finish the surface with Ni/Au or Cr/Au.

-branadic-

[The Soulman]

They do make and sell a lt1021H(ermetic), what does the bottled one bring to the table?

Edit: btw, A+ for effort   

[BravoV]

You should try another one but not in the glass bottle, maybe cast it in a "removable/disposable" casting container, and once the rosin is hardened, then dismantle it and put the whole "rosinated" chip in a stress free container.

The reason is I think rosin has quite high thermal expansion as I know this pretty well, old hardened violin rosin can crack if it had been through thermal cycling.

In the "stiff" glass bottle, thermal cycling could stress out mechanically the chip's body.

Worth an experiment.

My 2 cents.

[GerryBags]

You could rig up a humidity chamber with one of those piezo-misters hooked up to a controller and a very accurate hygrometer. You can dry air out quite effectively with perlite/vermiculite filters without altering its temperature, too. Just a thought.

[BravoV]

Curious what if you cool it down in freezer say overnight, once complete freezed, take it out and let it settles slowly to room temp and measure again.

Pretty sure thermal contraction will happen to the rosin even its hardened.

[BravoV]

The reason I'm interested with this rosin expansion/contraction is me & my dad inherited lots of these violin rosin blocks from my granddad. Remembered when I was a kid when my granddad was still around, when new never used, they're sort had the nice shiny convex surface like this casted block in the wood container.


Random image from Google.

Later on after few decades, when I saw them again, even unused, few already detached from the wood container still in whole piece, and the surface sort of concavely shaped, its like somehow it shrinked and loosing some volume throughout many years.

[BravoV]

Actually, if your objective to protect the IC from humidity and probably atmospheric pressure fluctuation, maybe you should try glassware work, as its relatively easy & cheap.

Seal the reference inside the glass tube, not exactly the same, but attaching this video to give you an idea.

Pretty sure this process will not toast the chip inside, as this video shows the liquid is not evaporated during the process.

https://youtu.be/GEmINGq8HDQ

[BravoV]

thanks BravoV,

where can i get that burner from? what fuel does it run? (butane?)

best regards.

-zia

Yep, those common butane gas torch works fine.

Just go to your local chemical lab equipment store, usually they have those glass tube in various diameters for chemical reaction, some are made that can be shaped easily.

One thing for sure, they're relatively cheap, and of course, practice or even sacrifice few of them for you to get use to handle it.

[BravoV]

Regarding the wires, I guess most reference you need only few, just heat the glass until it soft, and then use metal plier to sort of crimp the wires.

Again, you need to practice 1st before sacrificing your precious IC.

These are professionals, but to give you an idea on sealing the wires, watch the process at 2:30 ..

https://youtu.be/8n4WVRKkmww?t=2m29s

[BravoV]

meanwhile, i can choose a ref. for freezer testing as you said, what ref. would you prefer?
lt1021, lt1027, lt1036, lt1034, lm329bz, ltc6655 or lt6657 ?

I'm not a voltnut  , but if I were you ... I prefer LTZ1000. 

Seriously, just try perfecting your glass sealing technique, once you're comfortable with it, its just the matter of time you seal your ideal chip.

Once you sealed "captured" this little devil inside the glass, no need to worry about humidity or atmospheric fluctuations anymore, the only thing you need to manage is just the temperature.

[BravoV]

or you could use a couple of these (*very* cheap in pakistan)
you get 2 feed through's per device.

best regards.

-zia

Those neon bulb are too small in diameter in order to fit in the chip.

But the most important thing is, now you get it on sealing the chip inside the glass.

Just find a cheap butane torch, and start scavenging used small glass (thin wall) bottles that for pills/medicine and start experimenting. 

[BravoV]

@BravoV

you use the neonbulb as feedthrough stems only, get a bigger glas tube, seal these into that tube with connections to the chip in the bigger tube.

best regards.

-zia

Ok, I misunderstood you. I guess now you're thinking an advanced glassware work now. 

Looking forward to see your glassware creations here. 

[Gyro]

You may run into problems due to parasitic and reverse biased B-E junctions internal to the LT1021 if you disconnect the input supply with the battery on the output. Certainly an accidental short on the input would kill the LT1021.

Ideally you would have a series diode - obviously impractical on the output but I would certainly add one in series with the input.

Accidental loading of the output could cause the reference voltage to deviate for a long time too as the cells recharge.

[Gyro]

Ah yes, that would do it. Glad you've got it covered. 

[pigrew]

I found a random paper on Li-ion battery noise that suggests (if its results are to be believed) that a particular Li-ion battery has minimum voltage noise at around 3.5 V.

[pigrew]

I just read a little about NiCd and NiMH batteries. During charging, the voltage will tend to decrease once the battery reaches max charge. That, along with its relatively flat charge curve, maybe would make it control as a variable reference.

They are expensive, but perhaps the Murata UMAL/UMAC batteries would be best. They are a weird Lithium chemistry with a nominal voltage around 2.5V.

SLA lead acid batteries may also be a good choice? Or just go with a big capacitor?