Voltage Reference LT1021 CCN8-10 Best choice for Metrology applications

[Kiriakos GR]

Voltage Reference LT1021 this is single chip design aiming to deliver 10V at 10 mA.

Basically this is not just one product for every need and use.
Ultra stable DC this is used for powering things and not only as reference voltage.

They are application as for example in RF oscillators or in audio pre-amplifier in which stable DC this is highly important requirement.
In such applications one basic LT1021 with fixed voltage output ( Non adjustable) this will serve well, as this it would deliver better results than one generic DCV stabilizer chip of the family of :  7805, 7812 and similar.   

Best choice for Metrology applications this is LT1021 CCN8-10, this is a chip which I am also using for several years.
Specific LT1021 version this include  fine-adjust pin and better quality characteristics.

My choice at using  LT1021 CCN8-10 was and is as battery powered.
And I did that because my application this one Voltage reference source this able to challenge up to 7 1/2 digits ( High End ) bench-top multimeter.

Linear Technology this is a good company even about customers support, I did receive help from their engineers by email and they do care so their product this to deliver 100% satisfaction.

http://www.linear.com/product/LT1021

[Kiriakos GR]

Linear Technology it does a good job about product specifications but some times valuable information this seems hidden in a paragraph and application example which most customers will skip reading.

Few days ago I did rediscovered LT1021(product design)  by finding the information that a foam cover around this Voltage Reference this will further increase stability specification, because the foam will eliminate airflow.

I am using LT1021 this enclosed in to a larger project box this hosting my Seven dials resistance box project
There is no active cooling or passive cooling at my project, but I never thought to eliminate internal airflow and to totally isolate LT1021 package from even minimal airflow.

Today I did started working over this final step of improvement, and this is also an opportunity for ITTSB Blog visitors to see take-apart pictures of my own voltage reference. 

For now I will add few pictures of PCB I am using. 

 
   

[Kiriakos GR]

My foam sandwich this is now ready. 

Primary power supply I am using this this is a box including four Lithium batteries 18650 2200mAh 3,7V = 4,2 x4= 16,8
16,8 V this feeds one LM7812 this delivering stabilized 12,07V and also there is a small 47uF 16V filtering capacitor.

When I powered my LT1021 for a first time, output was at the negative side 9,9998 and I disliked that, so I did use LT1021 capability for fine adjust of the output by following the schematic.
By using as comparison point Fluke 28 and two U1272A plus the Hioki DT4282 and also the TIME ELECTRONICS UK 1048, I did succeed to touch 10V at five digits resolution.
I got little bit disappointed that my BM869 at 50V range this measures higher than expected and instead of a perfect 10.000 this measures 10.001 and at 500,000 counts mode this now measure 10,0015.

Currently I am working in a project of a Japanese software developer, and I helped him to improve his multimeter data-logging software so this to operates with fully adjustable Min - Max limits in the graph window.
The software it is compatible with several multimeter and Now fully compatible with BM869 at 500,000 counts mode.
The software has eight graticule resolution and last digit of BM869 at 500,000 counts mode can be logged for obtaining statistics of Voltage Reference stability.
Unfortunately for this experiment, BM869 seems to have approximately 180~200 micro-volt noise due product design.
Therefore LT1021 specifications cannot be verified until ITTSB Blog receive an superior bench-top multimeter.

Today my LT1021 with the foam sandwich and two hours warmup this looks very solid and I am very pleased so far. 
Regarding the new software, this is still under development, I will continue working as Beta tester and I will post final logging results of my LT1021 in a few days.       

[Kiriakos GR]

 
This is one of those times which a man feel proud of him self because he found by his own a NEW technical solution.

With the help of my Japanese software (high resolution data logger) I was finally able to log for several hours the behavior of my LT1021-10.
Several days of monitoring voltage and temperature it did payed out,  LT1021-10 at default circuit diagram (provided by LT) it is prone to drift due temperature change in the range 23~28 Celsius.

For LT1021-5 there is a suggestion about adding a single 1N4148 Zener diode from Trim pin to the center pin of the ADJ pot.
Such a solution will make the reference chip immune to drift due temperature.

LT1021-10 this is different than LT1021-5, addition of single Zener causing at 10K pot an increase of resistance, and the reference this is unable to be trimmed lower than 10.0012V (10V).
The solution which I thought and tested and works !!!!   This is adding quantity of five schottky diode SB160 in parallel.
With out any other addition at trim circuit, and by adding one filtering capacitor (Metallized Polyester Film Mylar Capacitor) at 800nF in the output, my reference chip become ROCK SOLID.    

I am currently data logging and monitoring the reference for 18H and the plan is to continue for another two days days.
Additionally I got as loan one Fluke 8246A so to verify the reference output with out drift which this caused by temperature change at portable multimeter. 

My ultimate target this is my reference to hit the mark of  1000Hours = 43 days of operation, and final trim correction will be performed at that time.

To be continued ...
   


   
 

[Kiriakos GR]

With some help from friends and their electronics lab at local industrial park, I did manage to smash 5 1/2 digits barrier.   

My LT1021 rocks the house at 6 1/2 digits  !! 

Fluke 8846A
DC range:   +/-  0.0024%  +0.0005