Ultra Precision Reference LTZ1000

[TiN]

I think condensed water melt and leaked

[cellularmitosis]

I wonder if you can alter the zener current enough to push the zero TC point over to -78C (dry ice temperature).  Very interesting!

[TiN]

Getting to DICE temps is no better than going to LN2 anyway. I don't miss using DICE, having cold acetone slurry and other nasty and messy stuff (used DICE few times to cool CPUs/GPUs). LN2/single-stage CC is so much easier once you solve air going into dewar.
Anyway, all of that is offtopic here. I'll try with heater on and LN2 again, once have more time.

[Galaxyrise]

I wonder if you can alter the zener current enough to push the zero TC point over to -78C (dry ice temperature).  Very interesting!

Remember that's not the curve for a LTZ at a stable current.

[chuckb]

There are a few techniques used to change the zero TC point of the LTZ1000.

Janaf and others experimented with this back around post #1144-1176 (2015)

Change the  Q2   Q1 collector resistor to 5.6M ohms or add a 20 ohm resistor between the top of the zener and the 7v circuit node. This resistor is identified by R60 on this forum. This was shown on the LTZ data sheet and used in the Datron 4910 Voltage reference. 
So increasing R2 from 68k to 5.6 Meg raised the Zero TC point (with unstabilized current) from -130 deg C to 40 deg C. The same happens with adding R60 and increasing its value from 0 to 20 ohms.

The R2 impact will be nonlinear but R60 is linear so we can do an estimate.

The goal of -78 deg C is 30% of the way between -130 deg C and 40 deg C, so you could try an R60 value of 6 ohms to minimize the basic temp co with dry ice.

I would shoot for a chip temp of -60 deg C and run the heater so you have 0.001 deg C chip stability. That would take an R60 value of 8 ohms.
This is all guesswork you understand!

For TiN a Q2 collector resistance of 39k like Datron would lower the Zero TC point to an unknown cooler temperature.

corrected a few values

[chuckb]

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.

TiN
You achieved very low noise here at 77K and 8 ma. Can you check the noise of this same LTZ setup under normal conditions (40 deg C, 4 ma)? The noise will of course be much higher at room temperature. But, how high? I would expect the noise increase caused by the current reduction (8 ma to 4 ma) to be 1.4 times. Then the remaining noise increase would be cause by the temperature change. It looks like the noise reduction from cooling may be substantial. I would like to quantify it.
What would life be like without numbers!

The only other place I have seen information on Zener noise vs temperature is in the LM399 data sheet. See attached.

Keep moving the State of the Art forward for us.

[TiN]

I tested 10mA before (different LTZ module and different chip) and reduction on noise was less than I expected, about 20% only. I think I posted data before.

[chuckb]

I tested 10mA before (different LTZ module and different chip) and reduction on noise was less than I expected, about 20% only. I think I posted data before.

I remember seeing that so I dug it up.

There was a question about overdriving LTZ with high current. Well, I got some practical data, using jumper chip as a specimen, since I don't care much about it's long-term stability.

The noise was evaluated with a 3458A which will not have the low noise floor like your dedicated tester. Or was there some other testing that I missed?

Also it would be best to evaluate the same chip that was tested at 77K so we know if the that Zener is SPECIAL.

[TiN]

I'll come back to more tests later on, the story of cryoLTZ not finished yet.

[chuckb]

Thermal resistance of the LTZ1000 non A package.

The data sheet has confusing (for me) information for the non A LTZ1000. The Electrical Characteristics table lists a thermal resistance of 80 deg K / watt. The last graph of the Typical Performance Characteristics indicates an 80 deg K rise with 0.5 watts of heater power or a thermal resistance of 160 deg K / watt. What is the correct value?

Has anybody performed testing to clarify this?
Thanks

[d-smes]

@chuckb-  I ran into that too.  See https://www.eevblog.com/forum/metrology/ultra-precision-reference-ltz1000/msg1043115/#msg1043115 and the next few responses.  Correct thermal resistance for the non-A part is about 160 K/W according to data sheet typical curves.   While I have not verified this, the 160 K/W compares nicely with the 150 K/W of other devices in a TO-5 package (e.g. LT1012) which would presumably have the same die attach and therefore similar thermal resistance as the non-A LTZ1000.   I suspect the 80 K/W in the data sheet is a typo.

[babysitter]

Damned, now a portion of the worst case happened - Dilligentminds removed posts, hurting quote-integrity of this thread. :-/

[BravoV]

Damned, now a portion of the worst case happened - Dilligentminds removed posts, hurting quote-integrity of this thread. :-/

Lesson learned, at least for me, use quote next time and save the posted images too if its important.

Also personally, I don't trust people who used external hosting for their images posted here in this forum, as I have the feeling that in their mind once they want to leave, all they need is just delete their posts and disable/terminate the images posted externally.

[TiN]

I don't trust people who used external hosting

.
Goes both ways, I don't trust putting data, photos and files to some oompa-lumpa web-hosting server nobody knows where and hoping it will be around in 10 years from now. 
If I decide to have my content withdrawn, it's still my content and right to do so.   

[BravoV]

I don't trust people who used external hosting

.
Goes both ways, I don't trust putting data, photos and files to some oompa-lumpa web-hosting server nobody knows where and hoping it will be around in 10 years from now. 

If I decide to have my content withdrawn, it's still my content and right to do so.   

Two ways ? Not always, as you're definitely with an agenda and surely not sincere sharing your posts here and still riding EEVBlog to gain popularity. That is one way. 

Just still curious why you haven't officially announce your own forum/web/portal/video blog ? I guess when the time comes one day, all your posts will suddenly gone from here, all images gone, hence its not sincere sharing as you're thinking you have the power with just one click to make it all gone from here.  Again, this is one way too.

Also probably that coincides with the timing of official launching of the "Tin Metrology" xdevs.com portal soon ? But good for you and your ambition, wish you luck. 

No doubt, all your posts is yours, but still I have the right to save local copy once you posted "publicly", and also I have the right to re-post in the internet too if someone needs it, without any catch, this, two ways.

[TiN]

Go ahead, save it all, more backups the better. Just don't forget credits and all, so other readers know the source originator.
I also lost you on "official announce" part, links are in my profile and under each post sub, how more official can it be. There is no official launching anything, that I know about either. 

Suggest we kill the offtopic on these good points and discuss LTZ designs instead.

[BravoV]

Agree, no more, I stopped here.

[ArthurDent]

I have started collecting precision resistors and other parts to build a LTZ1000 reference and ordered the KX boards from OSHPARK and would like to comment on the company.

When the boards arrived there was a problem and I emailed them and got an instant reply. They went above and beyond anything I expected to completely correct the problem. I am very impressed at their level of service and the quality of their produced boards.  I know occasionally we all hear about something we buy not living up to the claims about the product but I feel we also should comment on the great service and products we receive from companies like OSHPARK.  I’m impressed.

[d-smes]

OT

Dilligentminds removed posts, hurting quote-integrity of this thread. :-/

DiligentMinds contributed some great ideas and insights to this thread.   Does anyone have a backup of this thread from before his posts were removed?   I heard once you put something out on the internet, it's always out there... but where?

[hwj-d]

OT

Dilligentminds removed posts, hurting quote-integrity of this thread. :-/

DiligentMinds contributed some great ideas and insights to this thread.   Does anyone have a backup of this thread from before his posts were removed?   I heard once you put something out on the internet, it's always out there... but where?

OT:

Play arround a little with WaybackMachine
https://archive.org/web/
like this:
https://web.archive.org/web/20180217033926/https://www.eevblog.com/forum/index.php

or google cache.

€: Gradually one should start at suitable place in the forum its own thread with topic to Dilligentminds.

[Insatman]

I've laid out this LTZ1000 based 10V voltage reference board.   The circuit takes 18-24 VDC input (regulated) and produces a 10.0000 Volt output into high impedance.   This is an iteration of an earlier design so I know it works.  I'm trying to improve things a bit with the new layout and addition of some filtering.   

Some notes:
1)  The resistors with 4 terminals are just resistors with thru-holes set up for 0.150 or 0.4 inch spacing.  I'm using DipTrace for the layout  and by setting up the schematic this way it ensures I connect all the pads.
2)  The S1 connection is to separate the grounds so when I do the ground plane fill, those pads are not included.   I just install a jumper to connect them at a single point.
3)  The ferrites used are single turn intended for high-frequency noise suppression.   They are about the size of a 1/2W resistor and have near zero DC resistance.
4)  The board size is 2x3".  This is necessary to fit in my particular application.
5)  The space around the LTZ1000 will allow for a small Styrofoam hollow cube to be double-stick taped to the board providing some insulation and preventing air currents from affecting the LTZ1000
6)  The board will be housed in a cast aluminum enclosure lined with foam.

See attached pictures.

Anyone have a comments or suggestions before I commit to making these boards?

[Andreas]

Anyone have a comments or suggestions before I commit to making these boards?

- I would place D4 directly to the input to prevent C17 blowing in case of reverse polarity.
  Additional bonus: L1+L2 see directly a capacitor (C15) giving a PI filter.

- Is U4 really stable with 10nF load? -> I would add a "capacitive loading isolation cirquit"
  See R21+C22 in my cirquit (R21 has to be adapted to the OP-Amp type and may be up to 100 Ohms).

- I am missing the EMI capacitors at the most sensitive pins of the cirquit
  (C11+C12 in my design placed directly between base + emitter pins of the LTZ IC)

with best regards

Andreas

[Insatman]

Anyone have a comments or suggestions before I commit to making these boards?

- I would place D4 directly to the input to prevent C17 blowing in case of reverse polarity.
  Additional bonus: L1+L2 see directly a capacitor (C15) giving a PI filter.

- Is U4 really stable with 10nF load? -> I would add a "capacitive loading isolation cirquit"
  See R21+C22 in my cirquit (R21 has to be adapted to the OP-Amp type and may be up to 100 Ohms).

- I am missing the EMI capacitors at the most sensitive pins of the cirquit
  (C11+C12 in my design placed directly between base + emitter pins of the LTZ IC)

with best regards

Andreas

Moving the diode is a good idea.  I will try and modify the layout.

The OPA177FP op-amp is stable seeing 10nF directly, but a resistor will fit in the location of L3 just in case.  It's actually the same pad spacing.

I thought about putting caps there before but never saw it implemented, so I was reluctant.  I will see if I can fit them in the layout.  I'm assuming you are using film capacitors there?

Thanks...........Insatman

[Andreas]

I thought about putting caps there before but never saw it implemented, so I was reluctant.  I will see if I can fit them in the layout.  I'm assuming you are using film capacitors there?

Hello,

I have to admit that I am using X7R capacitors to get them really close to the pins.
(one is 0603 in the middle of the LTZ-pins the other 0805)
But of course film capacitors (e.g. SMD ECHU type) are also possible.

with best regards

Andreas

[Insatman]

Ok...I implemented all the suggestions.  On the new caps, I used two.  A 100nF film located some ways away from the LTZ1000 and an SMD 1206 33nF COG capacitor located very close to the chip on the underside of the board.   I happen to have quite a few of those in stock at the moment.   Both the board and schematic are updated.

Thanks Andreas.

Insatman.