LT1021 plastic

[GigaJoe]

im going cheap, as usual ...

a while ago i did 16 plastic LT1021, paralleling together over some balancing resistors, with adjustment  and inside a thermostat.
As the thermostat itself are good, all this construction became a bit better than disaster, an output swing at non predicted level, randomly going up and down.

Im goin to the next loop of cheap

so i pick random used LT1021-10 PDIP8 ,  add basic china multiturn resistor, 1N4148 as compensation, from +20C to +40C , i got negative swing of 4 microvolts. basically lucky. tossed in plastic container,  add 2 5W resistors, container filled by oil ,  everything in a thermos.   basic 12V regulator

it run it some period on March, (let say 500 hours) ; and on april I sporadically did a measurement , just to see a tendency ..  ; measurement device 34401a , with positive tempco, and room temp usually +18, and was +20  during mid April heatwave.

im guesin - 4 (or 6) such modules ,  adjusted to min tempco ,  a bit running a couple month ,then adjusted to the same output , (avoiding balancing resistors) probably make sense. 

attached file of sporadic measurement, and guessing Temp inside of oil bottle as thermistor not calibrated. 34401a I think +0.8PPM

[iMo]

Interesting.. Mind the measurement with 34401A is tricky, from my experience.. It needs 8-12h to fully stabilize and its TC depends on elapsed time after the power-on. A measurement here takes a day..
I collected around 40pcs of epoxy 1021-10 and -5, their TC -1 to +4ppm (simple finger method). The idea with the oil is interesting. Perhaps I would dry the epoxy first and then put the package(s) into the oil, the oil should not be hygroscopic.
Why do your temperature in the thermos flask vary so much (3C)?

[GigaJoe]

multimeter are constantly on , 2 years maybe ... switching display off . had some fun with insulate LM , add some shields, maximum possible filter caps, add some poly on top of tantalum. switch back\frond - soldered ;  kinda modding does it helps no idea.

the temperature not stabilized,  just 2 resistors ,  so it depend on ext. temp. and i think i stacked something on top , so temp dissipation does lower and temp, increased.
the lowest air room temp around +17 highest maybe 20-21 , in a month period.  I'm really have no idea why i add non stabilized heater,  guess it was assumption it smooth deviation, it probably can ,  but need carefully build a carrying case so excess power dissipation need more less constant.

im quite satisfied , had assumption it would much worse. LT1021 I actually pick with really high tempco like 20ppm ,  so ruthlessly fire it up , but it works,  I add 5 diodes to compensate tempco .  I have an idea to add diodes till temco change sign,  then add something ( potentiometer , something else)  to the last diode for adjusting the level how this diode will affect tempco.

[GigaJoe]

20230502 9.999843
20230806 9.999750

[iMo]

Yep, I still wonder how it is possible people so often report "great" results from their cheapo 1021 based calibrators..
Afaik Andreas and dietert1 reported good results when keeping the 1021 at elevated temperatures - like 50C and 42C, and 24x7 powered on.

[Andreas]

20230502 9.999843
20230806 9.999750

Hmm, I am missing at least temperature (and humidity for the plastic case).

Afaik Andreas

No I do not have a LT1021 under observation.
The only part (in SO-8) has about 7 ppm/K temperature drift.
So I use it for ratiometric (temperature) measurements.

As all buried zeners the LT1021 has a low ageing drift.
For the T.C. they should be temperature stabilized.
Against humidity one would use the metal can version.

with best regards

Andreas

[iMo]

Afaik Andreas

No I do not have a LT1021 under observation.

Ahh, my bad, I've thought those are the plastic 1021
https://www.eevblog.com/forum/metrology/small-oven-controler-for-voltage-reference/msg1655015/#msg1655015

[Andreas]

Hello,

No those are now mostly AD586LQ CERDIP8 after replacing the formerly LT1027 (DIP8).
Some MAX6250A, MAX6350, one LT1019, one REF102AP, one AD586MNZ (in PDIP)
and finally some LT1236AILS-8

with best regards

Andreas

[GigaJoe]

it cheap plastic ... random taken,  used 

it thermally compensated. by diodes , approx  -0.1-0.2 ppm
in oil bath , and heater , non-thermostatic,  around +45C
everything in  a plastic container, container in a metal thermos.
sooo.  basically humidity irrelevant, temp ...  maybe ... im curious how it goes  ... will see in next few month ...

[GigaJoe]

20230502 9.999843
20230806 9.999750
20240228 9.999792

no comments ... 

[magic]

As far as I understand, it should be possible to null linear TC of most LT1021 specimens with a pot between pins 1,8 with grounded wiper. Or fixed resistor(s). The external network will be in parallel with two factory trimmed on-chip resistors of a few kΩ each, so its resistance should be fairly high for low sensitivity. This adjustment affects output voltage too.

Maybe better than compensation with external diodes?

[iMo]

Do we know how the pins 1 and 8 are wired internally? (NC in the DS).
Perhaps this - https://www.eevblog.com/forum/metrology/decapping-the-lt1236ls8/msg2935094/#msg2935094

[Andreas]

Mhm,

does this work with LT1236 too?
The chips look very similar.

with best regards

Andreas

[magic]

Yes, this is what I meant.

Mhm,

does this work with LT1236 too?
The chips look very similar.

IIRC we couldn't find any difference between LT1021-5 and LT1236, other than different package and maybe different factory trimming.

LT1021-10 has the same adjustment network on pins 1,3,7,8 wired to the current mirror of its internal opamp.

[iMo]

But there are the 4 fuses (shorting the resistors) between pins 1/3/7/8, so it depends, whether it is possible or not..
Or, do blow up the fuses and then you may play with the balancing as you wish..
PS: I see only 2 fuses blown up in my 10 random 1021 samples (out of 37)..

[iMo]

I've set my PSU to 2V/50mA and I've tried to blow up the 4 fuses (all 4 were active/short in this 1021-5 sample).
Now I see aprox 2.5k\$\Omega\$ between pin8 and pin7 and between pin1 and pin3.
Still short between pin7 and gnd and between pin3 and gnd, however.

PS: With a little bit more voltage I blew up the 7-gnd and 3-gnd fuses as well.

8-7  aprox 2.57k\$\Omega\$
7-gnd   aprox 1.31k\$\Omega\$
1-3  aprox 2.55k\$\Omega\$
3-gnd  aprox 1.95k\$\Omega\$

8-gnd aprox 30.6mV
7-gnd aprox 10.5mV
3-gnd aprox 15.5mV
1-gnd aprox 34.4mV

[iMo]

FYI- With an 1k 10T WW pot wired between pin8 and pin1 (wiper to gnd) I can set the output voltage aprox -5mV..+4mV around the 5.000V. The TC is always positive in that range, perhaps a little bit lower in the "middle", however..
That is why most samples have their 4 fuses active (a short).