Ultra Precision Reference LTZ1000

[BU508A]

Board is the LTZ1048B including BMON (battery monitor and battery charger) and battery pack (12x AA NiMH) by Andreas.

I looked around here in the Forum, but I wasn't able to find something, which describes this LTZ1048B board in detail. 
Could you post a link please? Thanks.

Andreas

[branadic]

It's the board shown here and there is also some schematic within this thread. The BMON is described here.

-branadic-

[branadic]

A nice back can charm as well.

-branadic-

[TiN]

I thought it might be fun for someone to take a look on some random numbers.

It's HP A9 STD-reference, powered by different sources, different times in the year, with uncontrolled ambient conditions (typical +21 to +27c), measured by two different HP 3458A (calibrated off same source, though) and with Keithley 7168 JFET nanovolt scanner in between.



This very same reference I used March 2016 to calibrate my 3458A first time.



Original data, taken prior to shipping



Even with all involved, I'm impressed with the peak-peak number and data. 

Sample size is non-existent, but if one to judge from this single unit, then theory of having standard A9 3458A reference achieving ~2ppm/year stability might have some real possibility.

If one to count from first measurement after shipment (May 20, 2016), 7.184609 vs last measurement (January 11, 2018), 7.184636 VDC, change is +3.76ppm over 601 days, that is +2.3 ppm/year. Which is bold statement already, as own meters drift is on similar scale, so it's hard to tell who is who.

[TiN]

Well, I think I need lower noise low noise amplifier. Finally got around and started test I wanted to do for very long time. Here's some interesting data for discussion.

Test setup: MDO4054C scope + Pipelie's 1:10000 pre-amplifier from low-noise thread, so 1mV on amplifier output is equal to 100nV.
DUT:  Modified KX reference board, LTZ1000CH random chip. It's running with standard 120 ohm VHP100 and 75K PTF56 resistor.
Power supply for reference: Agilent E3649A, +15VDC.



Shown graph reveals ~270nV/pk-pk noise on 0.1Hz-10Hz band.
Output voltage is 7.0751391 V.

Own preamplifier + scope noise floor is around 110nV pk-pk.

[Kleinstein]

Having 260 nV noise from the reference + amplifier compared to 110 nV from just the amplifier is not a problem. To a good approximation one can compensate for the amplifiers own noise up to the point where the source noise is slightly lower than the amplifiers noise.
For looking at the noise data the RMS values are usually more practical than peak to peak values as these show much lower fluctuations.

[Magnificent Bastard]

Shown graph reveals ~270nV/pk-pk noise on 0.1Hz-10Hz band.
Output voltage is 7.0751391 V.
Own preamplifier + scope noise floor is around 110nV pk-pk.

I think you got very lucky, and have an LTZ1000 that has exceptionally low noise.  Better keep that one!

[TiN]

There is nothing special about this chip .

Kleinstein
Well, that is my concern. I want to bump zener current 4-6 times to get in ballpark of 100nV/pkpk noise. To be more confident in noise that would require <50 nV/pk-pk capable amplifier.

[TheSteve]

It feels very cold in this thread... 

[Andreas]

I think you got very lucky, and have an LTZ1000 that has exceptionally low noise.  Better keep that one!

Hello,

my crystal ball sees a smoking dewar with LN2 around the board.

Otherwise there would be something fishy with that LTZ1000.
The stray in noise is usually not that high.

with best regards

@Illya: photos please

Andreas

[CalMachine]

It feels very cold in this thread... 

Put some full-length pants on 

[TiN]

Hello,

my crystal ball sees a smoking dewar with LN2 around the board.
@Illya: photos please

Andreas

Yes, this is LTZ chip cooled down to 77K.
I don't have any worthy meters at work, just an agilent 34970A, so here is live log:
https://xdevs.com/cltz_test1/.
Its just an indication, as noise and stability in data just shows now good is LM399 in 34970A. LTZ chamber temperature measured with Type K TC and MAX31855 converter, tied to RPi by SPI bus.

Just checking if whole idea have any means to it, before investing into making 24/7 operation capable cryocooled LTZ reference. One of things I am not sure is how stable temperatures would stay, due to oxygen and air creep into dewar.
Heater drive on LTZ is disconnected ofc.

[Echo88]

Did you submerge the whole board in LN²? Since i know you have the money and are crazy enough to give it a thought: You could also buy a "Superconductor Technologies Superfilter" on ebay which contains a stirling-engine capable of producing LN² and can run indefinitely.

https://www.ebay.com/sch/i.html?_odkw=super+filter&_osacat=12576&_from=R40&_trksid=p2045573.m570.l1313.TR0.TRC0.H0.Xsuperfilter.TRS0&_nkw=superfilter&_sacat=12576
http://benkrasnow.blogspot.de/2008/08/diy-liquid-nitrogen-generator.html Here the stirling engine is shown and used.

Time to build that ultra low drift/noise 3458B. 

[Cerebus]

Did you submerge the whole board in LN²? Since i know you have the money and are crazy enough to give it a thought: You could also buy a "Superconductor Technologies Superfilter" on ebay which contains a stirling-engine capable of producing LN² and can run indefinitely.

One would just make a board that clamps directly to the cryocooler's cold finger and surround the whole assembly with a vacuum flask and evacuate it, that is, you can cut out the intermediary LN₂.

That just leaves the tiny problem of how to make a good low thermal emf connection across a ~230 K thermal difference.

[Echo88]

I dont know how stable you can regulate the cold-finger-temperature. Im just surprised that the whole LTZ-assembly still works at 77K and that the ref-voltage is still in the 7V-range.

[Cerebus]

I dont know how stable you can regulate the cold-finger-temperature.

Good question. I know it's good enough to not add noise to microwave amplifiers that are being cryocooled specifically to get low noise (That's where most of the 2nd hand cryocoolers come from).

[chuckb]

TiN
Great LN2 Experiment!
I extrapolated some temp control resistor values for you at the very cold temperatures. See attached PDF. This is based on the -55C to +125 deg C characterization that I did a few years ago.

Given a 7.0751 V zener -
For a -150 deg C control point the nominal 13k value for R4 should be 6.075k ohms. This would yield 1.0 V on pin 6 of the chip.
For a -175 deg C control point R4 would be 5.738k ohms. This would be 1.05 V on pin 6.

The 120 ohm resistor probably had 0.96 V across it so you would have been running 8 ma for this test. This is a guess because the Vbe of the Q1 does not match theory at this time. This is because of the internal Zener kelvin connections.

Good luck!

[TiN]

Cryocooler they say...



Not sure if she'll be alright, need to have my friend to test it. Anyone got software for these STI RX 850 boxes?
My biggest concern would be mechanical decoupling (these motors produce quite a vibration during operation) of LTZ reference module and insulation against environment (condensing air on the reference assembly wouldn't be desired experience).

chuckb
I doubt that LTZ curve stays linear on these extreme temperatures.
I can measure some points. Resistors and board are outside, at ambient temperature, only LTZ chip is chilled.

Also using LTZ's own oven likely to be pointless, as heat flux removal ability of liquid nitrogen is many hundreds watts per volume that I have here now.

[Echo88]

Maybe LN² between the coldfinger and LTZ to transport the heat to the coldfinger, while a steel mesh in the LN² would dampen vibrations sufficiently? And a regulated heater in the LN² to carefully regulate the LN²-temperature since the colfinger-temp will vary alot i assume.

[chuckb]

I was thinking of an earlier experiment where you put a pcb inside a copper cylinder and submerged it in Nitrogen. But I don't remember the details.

I would love to see the chip node voltages. For the temperature, remove the divider resistors and short the base (pin 6) to the collector ( pin 8 ) . The 70k resistor will be the 100ua current source.
 
Looks like the Lakeshore Silicon temp sensors stay linear down to around 30 deg K. See attached. Of course that will be an issue when you progress to Liquid Helium!

Running the heater should work fine. The Liquid Nitrogen is still isolated by the steel package. Using an LTZ1000A would work even better with the extra thermal and stress isolation of the special chip mounting. You will probably need the heater to get to 0.001 deg C temp stability and repeatability.

What is the control temperature of the cryocoolers?

Great experiment!

[mimmus78]

TiN, this is the right step to PJVS ...

Inviato dal mio ONEPLUS A5010 utilizzando Tapatalk

[Kleinstein]

The special feature of the LTZ1000 is the good heater and temperature reading. For a LN temperature reference is would be more something like the LTFLU / SZA263 - however I don't know if the intrinsic TC is still that low at LN temperature.

Anyway I son't think this is really practical, as there is quite some effort in LN und cooling / warming to room temperature is quite some stress for the parts.  Ideally one would get a JJ reference. At least squids are possible at LN temperature - so one could get a squid + coil based current reference. A squid can also make is nice low voltage amplifier and current sensor.

[TiN]

 

The special feature of the LTZ1000 is the good heater and temperature reading. For a LN temperature reference is would be more something like the LTFLU / SZA263 - however I don't know if the intrinsic TC is still that low at LN temperature.

I'd stick with LTZ, just because it's commercially available. Such experiments are likely to cause few chips to be dead/permanently shifted/corrupted anyway.

Anyway I son't think this is really practical, as there is quite some effort in LN und cooling / warming to room temperature is quite some stress for the parts.

As unpractical as having multiple 8.5-digit meters and more than 10 LTZ1000-based references at home. We use decaliters of LN2 daily in the lab, so it's no big deal to leave 30L dewar in corner running something for few weeks.

Ideally one would get a JJ reference. At least squids are possible at LN temperature - so one could get a squid + coil based current reference. A squid can also make is nice low voltage amplifier and current sensor.

Ideally. As soon these things go from mid-5-digit $$ costs to 4-digit, I'll break a bank and get one. Before that, we stuck with little unpractical experiments for fun.

[TiN]

LN2 slowly evaporated from the dewar over last week, so I've got interesting curve of LTZ tempco w/o heater (measured by 2001).
Interesting "island of stability" at around -130 C, before the tempco reverses.

Total 38MB log CSV with over 6 days of data.

Also chuckb's guess about

The 120 ohm resistor probably had 0.96 V across it so you would have been running 8 ma for this test.

is very correct, when chip was at LN2 temp, voltage over 120R was 0.963V .

[zhtoor]

LN2 slowly evaporated from the dewar over last week, so I've got interesting curve of LTZ tempco w/o heater (measured by 2001).
Interesting "island of stability" at around -130 C, before the tempco reverses.

Total 38MB log CSV with over 6 days of data.

Also chuckb's guess about

The 120 ohm resistor probably had 0.96 V across it so you would have been running 8 ma for this test.

is very correct, when chip was at LN2 temp, voltage over 120R was 0.963V .

hello,

what happened between -1 and +5 degC ?

regards.

-zia